Multi-object tracking (MOT) aims to track multiple objects while maintaining consistent identities across frames of a given video. In unmanned aerial vehicle (UAV) recorded videos, frequent viewpoint changes and complex UAV-ground relative motion dynamics pose significant challenges, which often lead to unstable affinity measurement and ambiguous association. Existing methods typically model motion and appearance cues separately, overlooking their spatio-temporal interplay and resulting in suboptimal tracking performance. In this work, we propose AMOT, which jointly exploits appearance and motion cues through two key components: an Appearance-Motion Consistency (AMC) matrix and a Motion-aware Track Continuation (MTC) module. Specifically, the AMC matrix computes bi-directional spatial consistency under the guidance of appearance features, enabling more reliable and context-aware identity association. The MTC module complements AMC by reactivating unmatched tracks through appearance-guided predictions that align with Kalman-based predictions, thereby reducing broken trajectories caused by missed detections. Extensive experiments on three UAV benchmarks, including VisDrone2019, UAVDT, and VT-MOT-UAV, demonstrate that our AMOT outperforms current state-of-the-art methods and generalizes well in a plug-and-play and training-free manner.

Vision Transformer (ViT) has shown impressive performance in image restoration due to its ability to capture a large receptive field. However, its complexity grows quadratically with input resolution, limiting its applicability for high-resolution images. In contrast, Convolutional Neural Networks (CNNs) are computationally efficient but are constrained by their inherently local receptive fields, which limit their ability to capture long-range pixel relationships. To address these challenges, we propose StarIR, which possesses the efficiency of CNNs while also capturing a large receptive field, similar to Transformers. StarIR incorporates two key innovations: 1) a dual-domain representation learning framework, with one branch processing spatial details and the other focusing on mesoscale interactions in the frequency domain; and 2) a high-dimensional feature fusion mechanism, the Star operation, which fuses information from both domains through element-wise multiplication, thereby enhancing representational capacity without increasing network width and depth. Our Star operation is followed by a channel attention unit to facilitate global feature modeling and enhance channel-wise interactions. Building on our straightforward yet powerful design principles, StarIR achieves state-of-the-art performance across 21 datasets covering six single-degradation image restoration tasks. Furthermore, our model performs favorably against leading algorithms in two all-in-one settings and demonstrates robustness on two composite-degradation datasets. In addition, StarIR extends well to several domain-specific applications, including ultra-high-definition (UHD) imaging, remote sensing, medical imaging, and underwater image enhancement.

3D mask presentation attack detection is crucial for protecting face recognition systems against the rising threat of 3D mask attacks. While most existing methods utilize multimodal features or remote photoplethysmography (rPPG) signals to distinguish between real faces and 3D masks, they face significant challenges, such as the high costs associated with multimodal sensors and limited generalization ability. Detection-related text descriptions offer concise, universal information and are cost-effective to obtain. However, the potential of vision-language multimodal features for 3D mask presentation attack detection remains unexplored. In this paper, we propose a novel knowledge-based prompt learning framework to explore the strong generalization capability of vision-language models for 3D mask presentation attack detection. Specifically, our approach incorporates entities and triples from knowledge graphs into the prompt learning process, generating fine-grained, task-specific explicit prompts that effectively harness the knowledge embedded in pre-trained vision-language models. Furthermore, considering different input images may emphasize distinct knowledge graph elements, we introduce a visual-specific knowledge filter based on an attention mechanism to refine relevant elements according to the visual context. Additionally, we leverage causal graph theory insights into the prompt learning process to further enhance the generalization ability of our method. During training, a spurious correlation elimination paradigm is employed, which removes category-irrelevant local image patches using guidance from knowledge-based text features, fostering the learning of generalized causal prompts that align with category-relevant local patches. Experimental results demonstrate that the proposed method achieves state-of-the-art intra- and cross-scenario detection performance on benchmark datasets.

The rapid advancement of large language models (LLMs) has intensified the need for effective mechanisms to transform continuous multimodal data into discrete representations suitable for language-based processing. Discrete tokenization, with vector quantization (VQ) as a central approach, offers both computational efficiency and compatibility with LLM architectures. Despite its growing importance, there is a lack of a comprehensive survey that systematically examines VQ techniques in the context of LLM-based systems. This work fills this gap by presenting the first structured taxonomy and analysis of discrete tokenization methods designed for LLMs. We categorize 8 representative VQ variants that span classical and modern paradigms and analyze their algorithmic principles, training dynamics, and integration challenges with LLM pipelines. Beyond algorithm-level investigation, we discuss existing research in terms of classical applications without LLMs, LLM-based single-modality systems, and LLM-based multimodal systems, highlighting how quantization strategies influence alignment, reasoning, and generation performance. In addition, we identify key challenges including codebook collapse, unstable gradient estimation, and modality-specific encoding constraints. Finally, we discuss emerging research directions such as dynamic and task-adaptive quantization, unified tokenization frameworks, and biologically inspired codebook learning. This survey bridges the gap between traditional vector quantization and modern LLM applications, serving as a foundational reference for the development of efficient and generalizable multimodal systems.

Crowd localization aims to predict the positions of humans in images of crowded scenes. While existing methods have made significant progress, two primary challenges remain: (i) a fixed number of evenly distributed anchors can cause excessive or insufficient predictions across regions in an image with varying crowd densities, and (ii) ranking inconsistency of predictions between the testing and training phases leads to the model being sub-optimal in inference. To address these issues, we propose a Consistency-Aware Anchor Pyramid Network (CAAPN) comprising two key components: an Adaptive Anchor Generator (AAG) and a Localizer with Augmented Matching (LAM). The AAG module adaptively generates anchors based on estimated crowd density in local regions to alleviate the anchor deficiency or excess problem. It also considers the spatial distribution prior to heads for better performance. The LAM module is designed to augment the predictions which are used to optimize the neural network during training by introducing an extra set of target candidates and correctly matching them to the ground truth. The proposed method achieves favorable performance against state-of-the-art approaches on five challenging datasets: ShanghaiTech A and B, UCF-QNRF, JHU-CROWD++, and NWPU-Crowd.

Given a single labeled examples, in-context segmentation aims to segment corresponding objects. This setting, known as one-shot segmentation in few-shot learning, explores the segmentation model's generalization ability and has been applied to various vision tasks, including scene understanding and image/video editing. While recent Segment Anything Models (SAMs) have achieved state-of-the-art results in interactive segmentation, these approaches are not directly applicable to in-context segmentation. In this work, we propose the Dual Consistency SAM (DC-SAM) method based on prompt-tuning to adapt SAM and SAM2 for in-context segmentation of both images and videos. Our key insights are to enhance the features of the SAM's prompt encoder in segmentation by providing high-quality visual prompts. When generating a mask prior from support images, we fuse the SAM features to better align the prompt encoder rather than relying solely on a pre-trained backbone. Then, we design a cycle-consistent cross-attention on fused features and initial visual prompts. This design leverages coarse masks from the SAM mask decoder to ensure consistency between features and visual prompts. Next, a dual-branch design is provided by using the discriminative positive and negative prompts in the prompt encoder. Furthermore, we design a simple mask-tube training strategy to adopt our proposed dual consistency method into the mask tube. Although the proposed DC-SAM is primarily designed for images, it can be seamlessly extended to the video domain with the support of SAM2. Given the absence of in-context segmentation in the video domain, we manually curate and construct the first benchmark from existing video segmentation datasets, namedIn-Context Video Object Segmentation (IC-VOS), to better assess the in-context capability of the model. Extensive experiments demonstrate that our method achieves 55.5 (+1.4) mIoU on COCO-20$^{i}$, 73.0 (+1.1) mIoU on PASCAL-5$^{i}$, and a$\mathcal {J\&F}$score of 71.52 on the proposed IC-VOS benchmark.

We present DrivingGaussian++, an efficient and effective framework for realistic reconstruction and controllable editing of surrounding dynamic autonomous driving scenes. DrivingGaussian++ models the static background with incremental 3D Gaussians and reconstructs moving objects with a composite dynamic Gaussian graph, ensuring accurate positions and occlusions. By integrating a LiDAR prior, it achieves detailed and consistent scene reconstruction, outperforming existing methods in dynamic scene reconstruction and photorealistic surround-view synthesis. DrivingGaussian++ supports training-free controllable editing for dynamic driving scenes, including texture modification, weather simulation, and object manipulation, leveraging multi-view images and depth priors. By integrating large language models (LLMs) and controllable editing, our method can automatically generate dynamic object motion trajectories and enhance their realism during the optimization process. DrivingGaussian++ demonstrates consistent and realistic editing results and generates dynamic multi-view driving scenarios, while significantly enhancing scene diversity.

Numerous diffusion models have been developed for 2D image synthesis and editing, and recently they are extended to 3D scene editing tasks. However, editing 3D scenes is still in its early stages, and the challenges of scene representations and multi-view consistency need to be addressed. A notable limitation of existing approaches is the need for specific modules for different edits and model retraining for each scene. To tackle these issues, we propose a novel and versatile text-driven 3D scene editing method, termed DN2N, which allows for the direct acquisition of the editing results without the requirement for retraining. Our method employs off-the-shelf text-based editing models of 2D images to modify the multi-view images of a 3D scene. A content filtering process is then applied to discard poorly edited images that disrupt 3D consistency. We consider the remaining inconsistency as a problem of removing noise perturbations and solve it by generating data with similar perturbation characteristics for training. We develop a versatile NeRF model structure and propose two novel cross-view regularization terms to help the DN2N mitigate these perturbations. Empirical results show that our method achieves multiple editing types based solely on text prompts, including but not limited to appearance editing, weather transition, object changing, and style transfer. Most importantly, DN2N exhibits a versatility of editing capabilities, eliminating the need to customize or retrain editing models for specific scenes or editing types. Namely, DN2N achieves comparable total editing time to the 3DGS-based editing method, enhancing its practical value.

Recent advancements in multimodal large language models (LLMs) have demonstrated significant potential across various domains, particularly in concept reasoning. However, their applications in understanding 3D environments remain limited, primarily offering textual or numerical outputs without generating dense, informative segmentation masks. This paper introduces Reason3D, a novel LLM designed for comprehensive 3D understanding. Reason3D processes point cloud data and text prompts to produce textual responses and segmentation masks, enabling advanced tasks such as 3D reasoning segmentation, hierarchical searching, express referring, and question answering with detailed mask outputs. We propose a hierarchical mask decoder that employs a coarse-to-fine approach to segment objects within expansive scenes. It begins with a coarse location estimation, followed by object mask estimation, using two unique tokens predicted by LLMs based on the textual query. Experimental results on large-scale ScanNet and Matterport3D datasets validate the effectiveness of our Reason3D across various tasks.

Large diffusion models demonstrate remarkable zeroshot capabilities in novel view synthesis from a single image. However, these models often face challenges in maintaining consistency across novel and reference views. A crucial factor leading to this issue is the limited utilization of contextual information from reference views. Specifically, when there is an overlap in the viewing frustum between two views, it is essential to ensure that the corresponding regions maintain consistency in both geometry and appearance. This observation leads to a simple yet effective approach, where we propose to use epipolar geometry to locate and retrieve overlapping information from the input view. This information is then incorporated into the generation of target views, eliminating the need for training or fine-tuning, as the process requires no learnable parameters. Furthermore, to enhance the overall consistency of generated views, we extend the utilization of epipolar attention to a multi-view setting, allowing retrieval of overlapping information from the input view and other target views. Qualitative and quantitative experimental results demonstrate the effectiveness of our method in significantly improving the consistency of synthesized views without the need for any fine-tuning. Moreover, This enhancement also boosts the performance of downstream applications such as 3D reconstruction. The code is available at https://github.com/botaoye/ConsisSyn.

Unsupervised federated learning for cross-modal retrieval has received increasing attention in recent years as it can free the requirement for annotations and avoid uploading original clients’ data to servers. Most existing methods focus on how to learn better local models and their aggregation to overcome data distribution drift across clients. Unlike prior works, we propose to address the data distribution problem by generating synthetic data, which can benefit existing federated learning methods. Specifically, we train a WGAN generator with three newly designed loss constraints on each client to improve the quality of the generated data. We first compute cluster prototypes to address the problem of lack of labels. Then, a direct contrastive loss between generated image and text features, an indirect contrastive loss with reference to cluster prototypes, and a Jensen-Shannon Divergence (JSD) loss also with reference to cluster prototypes work together to constrain the WGAN. The locally trained generators and local prototypes are sent to the server to generate and filter synthetic data with consideration of data distribution across all clients. The filtered data are used to train the aggregated global retrieval model, which is later sent to clients. The final global model becomes robust to all clients after several rounds of client-server iteration. Extensive experiments using four baselines across three datasets demonstrate that our method performs favourably against state-of-the-art methods.

Open-Vocabulary Semantic Segmentation (OVSS) has advanced with recent vision-language models (VLMs), enabling segmentation beyond predefined categories through various learning schemes. Notably, training-free methods offer scalable, easily deployable solutions for handling unseen data, a key goal of OVSS. Yet, a critical issue persists: lack of object-level context consideration when segmenting complex objects in the challenging environment of OVSS based on arbitrary query prompts. This oversight limits models’ ability to group semantically consistent elements within object and map them precisely to user-defined arbitrary classes. In this work, we introduce a novel approach that overcomes this limitation by incorporating object-level contextual knowledge within images. Specifically, our model enhances intra-object consistency by distilling spectral-driven features from vision foundation models into the attention mechanism of the visual encoder, enabling semantically coherent components to form a single object mask. Additionally, we refine the text embeddings with zero-shot object presence likelihood to ensure accurate alignment with the specific objects represented in the images. By leveraging object-level contextual knowledge, our proposed approach achieves state-of-the-art performance with strong generalizability across diverse datasets.

Image restoration aims to recover content from inputs degraded by various factors, such as adverse weather, blur, and noise. Perceptual Image Restoration (PIR) methods improve visual quality but often do not support downstream tasks effectively. On the other hand, Task-Oriented Image Restoration (TIR) methods focus on enhancing image utility for high-level vision tasks, sometimes compromising visual quality. This paper introduces UniRestore, a unified image restoration model that bridges the gap between PIR and TIR by using a diffusion prior. The diffusion prior is designed to generate images that align with human visual quality preferences, but these images are often unsuitable for TIR scenarios. To solve this limitation, UniRestore utilizes encoder features from an autoencoder to adapt the diffusion prior to specific tasks. We propose a Complementary Feature Restoration Module (CFRM) to reconstruct degraded encoder features and a Task Feature Adapter (TFA) module to facilitate adaptive feature fusion in the decoder. This design allows UniRestore to optimize images for both human perception and downstream task requirements, addressing discrepancies between visual quality and functional needs. Integrating these modules also enhances UniRestore’s adaptability and efficiency across diverse tasks. Extensive experiments demonstrate the superior performance of UniRestore in both PIR and TIR scenarios.

Video personalization methods allow us to synthesize videos with specific concepts such as people, pets, and places. However, existing methods often focus on limited domains, require time-consuming optimization per subject, or support only a single subject. We present Video Alchemist—a video model with built-in multi-subject, openset personalization capabilities for both foreground objects and background, eliminating the need for time-consuming test-time optimization. Our model is built on a new Diffusion Transformer module that fuses each conditional reference image and its corresponding subject-level text prompt with cross-attention layers. Developing such a large model presents two main challenges: dataset and evaluation. First, as paired datasets of reference images and videos are extremely hard to collect, we sample selected video frames as reference images and synthesize a clip of the target video. However, while models can easily denoise training videos given reference frames, they fail to generalize to new contexts. To mitigate this issue, we design a new automatic data construction pipeline with extensive image augmentations. Second, evaluating open-set video personalization is a challenge in itself. To address this, we introduce a personalization benchmark that focuses on accurate subject fidelity and supports diverse personalization scenarios. Finally, our extensive experiments show that our method significantly outperforms existing personalization methods in both quantitative and qualitative evaluations.

Generating realistic human videos remains a challenging task, with the most effective methods currently relying on a human motion sequence as a control signal. Existing approaches often use existing motion extracted from other videos, which restricts applications to specific motion types and global scene matching. We propose Move-in-2D, a novel approach to generate human motion sequences conditioned on a scene image, allowing for diverse motion that adapts to different scenes. Our approach utilizes a diffusion model that accepts both a scene image and text prompt as inputs, producing a motion sequence tailored to the scene. To train this model, we collect a large-scale video dataset featuring single-human activities, annotating each video with the corresponding human motion as the target output. Experiments demonstrate that our method effectively predicts human motion that aligns with the scene image after projection. Furthermore, we show that the generated motion sequence improves human motion quality in video synthesis tasks.

Convolutional neural networks (CNNs) and Vision Transformers (ViTs) have achieved excellent performance in image restoration. While ViTs generally outperform CNNs by effectively capturing long-range dependencies and input-specific characteristics, their computational complexity increases quadratically with image resolution. This limitation hampers their practical application in high-resolution image restoration. In this paper, we propose a simple yet effective visual state space model (EVSSM) for image deblurring, leveraging the benefits of state space models (SSMs) for visual data. In contrast to existing methods that employ several fixed-direction scanning for feature extraction, which significantly increases the computational cost, we develop an efficient visual scan block that applies various geometric transformations before each SSM-based module, capturing useful non-local information and maintaining high efficiency. In addition, to more effectively capture and represent local information, we propose an efficient discriminative frequency domain-based feedforward network (EDFFN), which can effectively estimate useful frequency information for latent clear image restoration. Extensive experimental results show that the proposed EVSSM performs favorably against state-of-the-art methods on benchmark datasets and real-world images.

The goal of image cropping is to identify visually appealing crops in an image. Conventional methods are trained on specific datasets and fail to adapt to new requirements. Recent breakthroughs in large vision-language models (VLMs) enable visual in-context learning without explicit training. However, downstream tasks with VLMs remain under explored. In this paper, we propose an effective approach to leverage VLMs for image cropping. First, we propose an efficient prompt retrieval mechanism for image cropping to automate the selection of in-context examples. Second, we introduce an iterative refinement strategy to iteratively enhance the predicted crops. The proposed framework, we refer to as Cropper, is applicable to a wide range of cropping tasks, including free-form cropping, subject-aware cropping, and aspect ratio-aware cropping. Extensive experiments demonstrate that Cropper significantly outperforms state-of-the-art methods across several benchmarks.

Aligning text-to-image (T2I) diffusion models with preference optimization is valuable for human-annotated datasets, but the heavy cost of manual data collection limits scalability. Using reward models offers an alternative, however, current preference optimization methods fall short in exploiting the rich information, as they only consider pairwise preference distribution. Furthermore, they lack generalization to multi-preference scenarios and struggle to handle inconsistencies between rewards. To address this, we present Calibrated Preference Optimization (CaPO), a novel method to align T2I diffusion models by incorporating the general preference from multiple reward models without human annotated data. The core of our approach involves a reward calibration method to approximate the general preference by computing the expected win-rate against the samples generated by the pretrained models. Additionally, we propose a frontier-based pair selection method that effectively manages the multi-preference distribution by selecting pairs from Pareto frontiers. Finally, we use regression loss to fine-tune diffusion models to match the difference between calibrated rewards of a selected pair. Experimental results show that CaPO consistently outperforms prior methods, such as Direct Preference Optimization (DPO), in both single and multi-reward settings validated by evaluation on T2I benchmarks, including GenEval and T2I-Compbench.

The increasing demand for immersive AR/VR applications and spatial intelligence has heightened the need to generate high-quality scene-level and 360° panoramic video. However, most video diffusion models are constrained by limited resolution and aspect ratio, which restricts their applicability to scene-level dynamic content synthesis. In this work, we propose DynamicScaler, addressing these challenges by enabling spatially scalable and panoramic dynamic scene synthesis that preserves coherence across panoramic scenes of arbitrary size. Specifically, we introduce a Offset Shifting Denoiser, facilitating efficient, synchronous, and coherent denoising panoramic dynamic scenes via a diffusion model with fixed resolution through a seamless rotating Window, which ensures seamless boundary transitions and consistency across the entire panoramic space, accommodating varying resolutions and aspect ratios. Additionally, we employ a Global Motion Guidance mechanism to ensure both local detail fidelity and global motion continuity. Extensive experiments demonstrate our method achieves superior content and motion quality in panoramic scene-level video generation, offering a training-free, efficient, and scalable solution for immersive dynamic scene creation with constant VRAM consumption regardless of the output video resolution. Project page is available at https://dynamicscaler.pages.dev/new.

We present a unified network for simultaneously generating videos and their corresponding entity segmentation and depth maps from text prompts. We utilize colormap to represent entity masks and depth maps, tightly integrating dense prediction with RGB video generation. Introducing dense prediction information improves video generation’s consistency and motion smoothness without increasing computational costs. Incorporating learnable task embeddings brings multiple dense prediction tasks into a single model, enhancing flexibility and further boosting performance. We further propose a large-scale dense prediction video dataset Panda-Dense, addressing the issue that existing datasets do not concurrently contain captions, videos, segmentation, or depth maps. Comprehensive experiments demonstrate the high efficiency of our method, surpassing the state-of-the-art in terms of video quality, consistency, and motion smoothness.

Grounding natural language queries in graphical user interfaces (GUIs) poses unique challenges due to the diversity of visual elements, spatial clutter, and the ambiguity of language. In this paper, we introduce DiMo-GUI, a training-free framework for GUI grounding that leverages two core strategies: dynamic visual grounding and modality-aware optimization. Instead of treating the GUI as a monolithic image, our method splits the input into textual elements and iconic elements, allowing the model to reason over each modality independently using general-purpose vision-language models. When predictions are ambiguous or incorrect, DiMo-GUI dynamically focuses attention by generating candidate focal regions centered on the model’s initial predictions and incrementally zooms into subregions to refine the grounding result. This hierarchical refinement process helps disambiguate visually crowded layouts without the need for additional training or annotations. We evaluate our approach on standard GUI grounding benchmarks and demonstrate consistent improvements over baseline inference pipelines, highlighting the effectiveness of combining modality separation with region-focused reasoning.

Multi-modal Large Language Models (MLLMs) have dramatically advanced the research field and delivered powerful vision-language understanding capabilities. However, these models often inherit deep-rooted social biases from their training data, leading to uncomfortable responses with respect to attributes such as race and gender. This paper addresses the issue of social biases in MLLMs by i) introducing a comprehensive counterfactual dataset with multiple social concepts (CMSC), which complements existing datasets by providing 18 diverse and balanced social concepts; and ii) proposing a counter-stereotype debiasing (CSD) strategy that mitigates social biases in MLLMs by leveraging the opposites of prevalent stereotypes. CSD incorporates both a novel bias-aware data sampling method and a loss rescaling method, enabling the model to effectively reduce biases. We conduct extensive experiments with four prevalent MLLM architectures. The results demonstrate the advantage of the CMSC dataset and the edge of CSD strategy in reducing social biases compared to existing competing methods, without compromising the overall performance on general multi-modal reasoning benchmarks.

Unpaired image dehazing has attracted increasing attention due to its flexible data requirements during model training. Dominant methods based on contrastive learning not only introduce haze-unrelated content information, but also ignore haze-specific properties in the frequency domain (\ie,~haze-related degradation is mainly manifested in the amplitude spectrum). To address these issues, we propose a novel frequency domain-based diffusion model, named \ours, for fully exploiting the beneficial knowledge in unpaired clear data. In particular, inspired by the strong generative ability shown by Diffusion Models (DMs), we tackle the dehazing task from the perspective of frequency domain reconstruction and perform the DMs to yield the amplitude spectrum consistent with the distribution of clear images. To implement it, we propose an Amplitude Residual Encoder (ARE) to extract the amplitude residuals, which effectively compensates for the amplitude gap from the hazy to clear domains, as well as provide supervision for the DMs training. In addition, we propose a Phase Correction Module (PCM) to eliminate artifacts by further refining the phase spectrum during dehazing with a simple attention mechanism. Experimental results demonstrate that our \ours outperforms other state-of-the-art methods on both synthetic and real-world datasets.

Estimating geometry from dynamic scenes, where objects move and deform over time, remains a core challenge in computer vision. Current approaches often rely on multi-stage pipelines or global optimizations that decompose the problem into subtasks, like depth and flow, leading to complex systems prone to errors. In this paper, we present Motion DUSt3R (MonST3R), a novel geometry-first approach that directly estimates per-timestep geometry from dynamic scenes. Our key insight is that by simply estimating a pointmap for each timestep, we can effectively adapt DUSt3R’s representation, previously only used for static scenes, to dynamic scenes. However, this approach presents a significant challenge: the scarcity of suitable training data, namely dynamic, posed videos with depth labels. Despite this, we show that by posing the problem as a fine-tuning task, identifying several suitable datasets, and strategically training the model on this limited data, we can surprisingly enable the model to handle dynamics, even without an explicit motion representation. Based on this, we introduce new optimizations for several downstream video-specific tasks and demonstrate strong performance on video depth and camera pose estimation, outperforming prior work in terms of robustness and efficiency. Moreover, MonST3R shows promising results for primarily feed-forward 4D reconstruction. Interactive 4D results, source code, and trained models are available at: https://monst3r-project.github.io/.

Tracking and segmenting multiple similar objects with distinct or complex parts in long-term videos is particularly challenging due to the ambiguity in identifying target components and the confusion caused by occlusion, background clutter, and changes in appearance or environment over time. In this paper, we propose a robust video object segmentation framework that learns spatial-semantic features and discriminative object queries to address the above issues. Specifically, we construct a spatial-semantic block comprising a semantic embedding component and a spatial dependency modeling part for associating global semantic features and local spatial features, providing a comprehensive target representation. In addition, we develop a masked cross-attention module to generate object queries that focus on the most discriminative parts of target objects during query propagation, alleviating noise accumulation to ensure effective long-term query propagation. The experimental results show that the proposed method sets new state-of-the-art performance on multiple data sets, including the DAVIS2017 test (\textbf{87.8\%}), YoutubeVOS 2019 (\textbf{88.1\%}), MOSE val (\textbf{74.0\%}), and LVOS test (\textbf{73.0\%}), which demonstrate the effectiveness and generalization capacity of the proposed method. We will make all the source code and trained models publicly available.

We introduce \textbf{OmnixR}, an evaluation suite designed to benchmark state-of-the-art Omni-modality Language Models (OLMs), such as GPT-4o and Gemini. Evaluating OLMs, which integrate multiple modalities such as text, vision, and audio, presents unique challenges. Particularly, the user message might often consist of multiple modalities, such that OLMs have to establish holistic understanding and reasoning across modalities to accomplish the task. Existing benchmarks are limited to single-modality or dual-modality tasks (e.g., image+text or video+text), overlooking comprehensive multi-modal assessments of model reasoning. To address this, OmnixR offers two evaluation variants: (1) OmnixR-synth: a synthetic dataset generated automatically by translating text into multiple modalities—audio, images, video, and hybrids Omnify!. (2) OmnixR-real: a real-world dataset, manually curated and annotated by experts, for evaluating cross-modal reasoning in natural settings. OmnixR presents a unique evaluation towards assessing OLMs over a diverse mix of modalities, such as a question that involves video, audio, and text, providing a rigorous cross-modal reasoning testbed than any existing benchmarks. Our experiments find that all state-of-the-art OLMs struggles with OmnixR questions that require integrating information from multiple modalities to answer. Further analysis highlight differences in reasoning behavior and underscoring the challenges of omni-modal AI alignment.

Recent progress in image generation has sparked research into controlling these models through condition signals, with various methods addressing specific challenges in conditional generation. Instead of proposing another specialized technique, we introduce a simple, unified framework to handle diverse conditional generation tasks involving a specific image-condition correlation. By learning a joint distribution over a correlated image pair (e.g. image and depth) with a diffusion model, our approach enables versatile capabilities via different inference-time sampling schemes, including controllable image generation (e.g. depth to image), estimation (e.g. image to depth), signal guidance, joint generation (image \& depth), and coarse control. Previous attempts at unification often introduce complexity through multi-stage training, architectural modification, or increased parameter counts. In contrast, our simplified formulation requires a single, computationally efficient training stage, maintains the standard model input, and adds minimal learned parameters (15% of the base model). Moreover, our model supports additional capabilities like non-spatially aligned and coarse conditioning. Extensive results show that our single model can produce comparable results with specialized methods and better results than prior unified methods. We also demonstrate that multiple models can be effectively combined for multi-signal conditional generation.

3D Gaussian Splatting (3DGS) has shown promising results for Novel View Synthesis. However, while it is quite effective when based on high-quality images, its performance declines as image quality degrades, due to lack of resolution, motion blur, noise, compression artifacts, or other factors common in real-world data collection. While some solutions have been proposed for specific types of degradation, general techniques are still missing. To address the problem, we propose a robust HQGS that significantly enhances the 3DGS under various degradation scenarios. We first analyze that 3DGS lacks sufficient attention in some detailed regions in low-quality scenes, leading to the absence of Gaussian primitives in those areas and resulting in loss of detail in the rendered images. To address this issue, we focus on leveraging edge structural information to provide additional guidance for 3DGS, enhancing its robustness. First, we introduce an edge-semantic fusion guidance module that combines rich texture information from high-frequency edge-aware maps with semantic information from images. The fused features serve as prior guidance to capture detailed distribution across different regions, bringing more attention to areas with detailed edge information and allowing for a higher concentration of Gaussian primitives to be assigned to such areas. Additionally, we present a structural cosine similarity loss to complement pixel-level constraints, further improving the quality of the rendered images. Extensive experiments demonstrate that our method offers better robustness and achieves the best results across various degraded scenes. Source code and trained models are publicly available at: \url{https://github.com/linxin0/HQGS}.

Recent segmentation methods, which adopt large-scale data training and transformer architecture, aim to create one foundation model that can perform multiple tasks. However, most of these methods rely on heavy encoder and decoder frameworks, hindering their performance in real-time scenarios. To explore real-time segmentation, recent advancements primarily focus on semantic segmentation within specific environments, such as autonomous driving. However, they often overlook the generalization ability of these models across diverse scenarios. Therefore, to fill this gap, this work explores a novel real-time segmentation setting called real-time multi-purpose segmentation. It contains three fundamental sub-tasks: interactive segmentation, panoptic segmentation, and video instance segmentation. Unlike previous methods, which use a specific design for each task, we aim to use only a single end-to-end model to accomplish all these tasks in real-time. To meet real-time requirements and balance multi-task learning, we present a novel dynamic convolution-based method, Real-Time Multi-Purpose SAM (RMP-SAM). It contains an efficient encoder and an efficient decoupled adapter to perform prompt-driven decoding. Moreover, we further explore different training strategies and one new adapter design to boost co-training performance further. We benchmark several strong baselines by extending existing works to support our multi-purpose segmentation. Extensive experiments demonstrate that RMP-SAM is effective and generalizes well on proposed benchmarks and other specific semantic tasks. Our implementation of RMP-SAM achieves the optimal balance between accuracy and speed for these tasks. The code is released at \url{https://github.com/xushilin1/RAP-SAM}

We introduce NoPoSplat, a feed-forward model capable of reconstructing 3D scenes parameterized by 3D Gaussians from unposed sparse multi-view images. Our model, trained exclusively with photometric loss, achieves real-time 3D Gaussian reconstruction during inference. To eliminate the need for accurate pose input during reconstruction, we anchor one input view's local camera coordinates as the canonical space and train the network to predict Gaussian primitives for all views within this space. This approach obviates the need to transform Gaussian primitives from local coordinates into a global coordinate system, thus avoiding errors associated with per-frame Gaussians and pose estimation. To resolve scale ambiguity, we design and compare various intrinsic embedding methods, ultimately opting to convert camera intrinsics into a token embedding and concatenate it with image tokens as input to the model, enabling accurate scene scale prediction. We utilize the reconstructed 3D Gaussians for novel view synthesis and pose estimation tasks and propose a two-stage coarse-to-fine pipeline for accurate pose estimation. Experimental results demonstrate that our pose-free approach can achieve superior novel view synthesis quality compared to pose-required methods, particularly in scenarios with limited input image overlap. For pose estimation, our method, trained without ground truth depth or explicit matching loss, significantly outperforms the state-of-the-art methods with substantial improvements. This work makes significant advances in pose-free generalizable 3D reconstruction and demonstrates its applicability to real-world scenarios. Code and trained models are available at https://noposplat.github.io/.

Recent advances in vision-language models (VLMs) have made significant progress in downstream tasks that require quantitative concepts such as facial age estimation and image quality assessment, enabling VLMs to explore applications like image ranking and retrieval. However, existing studies typically focus on the reasoning based on a single image and heavily depend on text prompting, limiting their ability to learn comprehensive understanding from multiple images. To address this, we propose an effective yet efficient approach that reframes the CLIP model into a learning-to-rank task and introduces a lightweight adapter to augment CLIP for text-guided image ranking. Specifically, our approach incorporates learnable prompts to adapt to new instructions for ranking purposes and an auxiliary branch with ranking-aware attention, leveraging text-conditioned visual differences for additional supervision in image ranking. Our ranking-aware adapter consistently outperforms fine-tuned CLIPs on various tasks and achieves competitive results compared to state-of-the-art models designed for specific tasks like facial age estimation and image quality assessment. Overall, our approach primarily focuses on ranking images with a single instruction, which provides a natural and generalized way of learning from visual differences across images, bypassing the need for extensive text prompts tailored to individual tasks.

While recent low-cost radar-camera approaches have shown promising results in multi-modal 3D object detection, both sensors face challenges from environmen- tal and intrinsic disturbances. Poor lighting or adverse weather conditions de- grade camera performance, while radar suffers from noise and positional ambigu- ity. Achieving robust radar-camera 3D object detection requires consistent perfor- mance across varying conditions, a topic that has not yet been fully explored. In this work, we first conduct a systematic analysis of robustness in radar-camera de- tection on five kinds of noises and propose RobuRCDet, a robust object detection model in bird’s eye view (BEV). Specifically, we design a 3D Gaussian Expan- sion (3DGE) module to mitigate inaccuracies in radar points, including position, Radar Cross-Section (RCS), and velocity. The 3DGE uses RCS and velocity priors to generate a deformable kernel map and variance for kernel size adjustment and value distribution. Additionally, we introduce a weather-adaptive fusion module, which adaptively fuses radar and camera features based on camera signal confi- dence. Extensive experiments on the popular benchmark, nuScenes, show that our RobuRCDet achieves competitive results in regular and noisy conditions. The source codes and trained models will be made available.

We present Layout-Your-3D, a framework that allows controllable and compositional 3D generation from text prompts. Existing text-to-3D methods often struggle to generate assets with plausible object interactions or require tedious optimization processes. To address these challenges, our approach leverages 2D layouts as a blueprint to facilitate precise and plausible control over 3D generation. Starting with a 2D layout provided by a user or generated from a text description, we first create a coarse 3D scene using a carefully designed initialization process based on efficient reconstruction models. To enforce coherent global 3D layouts and enhance the quality of instance appearances, we propose a collision-aware layout optimization process followed by instance-wise refinement. Experimental results demonstrate that Layout-Your-3D yields more reasonable and visually appealing compositional 3D assets while significantly reducing the time required for each prompt. Additionally, Layout-Your-3D can be easily applicable to downstream tasks, such as 3D editing and object insertion.

With the growing interest in embodied and spatial intelligence, accurately predicting trajectories in 3D environments has become increasingly critical. However, no datasets have been explicitly designed to study 3D trajectory prediction. To this end, we contribute a 3D motion trajectory (3DMoTraj) dataset collected from unmanned underwater vehicles (UUVs) operating in oceanic environments. Mathematically, trajectory prediction becomes significantly more complex when transitioning from 2D to 3D. To tackle this challenge, we analyze the prediction complexity of 3D trajectories and propose a new method consisting of two key components: decoupled trajectory prediction and correlated trajectory refinement. The former decouples inter-axis correlations, thereby reducing prediction complexity and generating coarse predictions. The latter refines the coarse predictions by modeling their inter-axis correlations. Extensive experiments show that our method significantly improves 3D trajectory prediction accuracy and outperforms state-of-the-art methods. Both the 3DMoTraj dataset and the method are available at https://github.com/zhouhao94/3DMoTraj.

With the widespread adoption of wearable devices, human activity recognition (HAR), which is an essential branch of multivariate time series classification (MTSC), has been broadly applied in areas such as intelligent health monitoring, smart homes, and sports training. Traditional time series processing methods focus on time-domain features and often neglect frequency-domain characteristics; however, much of the critical information related to human activities is usually reflected in the frequency components of signals. To address this issue, we propose a Fourier transform convolution fusion module for extracting frequency-domain features and integrating them with time-domain features. Based on this module, we further propose an innovative Fourier transform convolution network (FTCNet). FTC-Net efficiently integrates time-domain and frequencydomain features by combining the Fourier transform convolution fusion module with an efficient channel attention block (ECA block), which enhances the model's ability to extract key features. The experimental results show that FTC-Net achieved excellent classification performance across 11 HAR-related datasets, with an average accuracy of $87.85 \%$, significantly outperforming current mainstream models. The results of ablation studies also demonstrate the positive impact of frequency-domain features on model performance. FTC-Net offers a new solution for complex human activity recognition tasks and opens new research directions for integrating time-domain and frequency-domain features.

Motion is one of the key components in deformable 3D scenes. Generative video models allow users to animate static scenes with text prompts for novel motion, but when it comes to 4D reconstruction, such reanimations often fall apart. The generated videos often suffer from geometric artifacts, implausible motion, and occlusions, which hinder physically consistent 4D reanimation. In this work, we introduce \textbf{Restage4D}, a geometry-preserving pipeline for deformable scene reconstruction from a single edited video. Our key insight is to leverage the unedited original video as an additional source of supervision, allowing the model to propagate accurate structure into occluded and disoccluded regions. To achieve this, we propose a video-rewinding training scheme that temporally bridges the edited and original sequences via a shared motion representation. We further introduce an occlusion-aware ARAP regularization to preserve local rigidity, and a disocclusion backtracing mechanism that supplements missing geometry in the canonical space. Together, these components enable robust reconstruction even when the edited input contains hallucinated content or inconsistent motion. We validate Restage4D on DAVIS and PointOdyssey, demonstrating improved geometry consistency, motion quality, and 3D tracking performance. Our method not only preserves deformable structure under novel motion, but also automatically corrects errors introduced by generative models, bridging the gap between flexible video synthesis and physically grounded 4D reconstruction.

Although diffusion-based models can generate high-quality and high-resolution video sequences from textual or image inputs, they lack explicit integration of geometric cues when controlling scene lighting and visual appearance across frames. To address this limitation, we propose IllumiCraft, an end-to-end diffusion framework accepting three complementary inputs: (1) high-dynamic-range (HDR) video maps for detailed lighting control; (2) synthetically relit frames with randomized illumination changes (optionally paired with a static background reference image) to provide appearance cues; and (3) 3D point tracks that capture precise 3D geometry information. By integrating the lighting, appearance, and geometry cues within a unified diffusion architecture, IllumiCraft generates temporally coherent videos aligned with user-defined prompts. It supports the background-conditioned and text-conditioned video relighting and provides better fidelity than existing controllable video generation methods.

We introduce the Deformable Gaussian Splats Large Reconstruction Model (DGS-LRM), the first feed-forward method predicting deformable 3D Gaussian splats from a monocular posed video of any dynamic scene. Feed-forward scene reconstruction has gained significant attention for its ability to rapidly create digital replicas of real-world environments. However, most existing models are limited to static scenes and fail to reconstruct the motion of moving objects. Developing a feed-forward model for dynamic scene reconstruction poses significant challenges, including the scarcity of training data and the need for appropriate 3D representations and training paradigms. To address these challenges, we introduce several key technical contributions: an enhanced large-scale synthetic dataset with ground-truth multi-view videos and dense 3D scene flow supervision; a per-pixel deformable 3D Gaussian representation that is easy to learn, supports high-quality dynamic view synthesis, and enables long-range 3D tracking; and a large transformer network that achieves real-time, generalizable dynamic scene reconstruction. Extensive qualitative and quantitative experiments demonstrate that DGS-LRM achieves dynamic scene reconstruction quality comparable to optimization-based methods, while significantly outperforming the state-of-the-art predictive dynamic reconstruction method on real-world examples. Its predicted physically grounded 3D deformation is accurate and can be readily adapted for long-range 3D tracking tasks, achieving performance on par with state-of-the-art monocular video 3D tracking methods.

This paper addresses the scarcity of low-cost but high-dexterity platforms for collecting real-world multi-fingered robot manipulation data towards generalist robot autonomy. To achieve it, we propose the RAPID Hand, a co-optimized hardware and software platform where the compact 20-DoF hand, robust whole-hand perception, and high-DoF teleoperation interface are jointly designed. Specifically, RAPID Hand adopts a compact and practical hand ontology and a hardware-level perception framework that stably integrates wrist-mounted vision, fingertip tactile sensing, and proprioception with sub-7 ms latency and spatial alignment. Collecting high-quality demonstrations on high-DoF hands is challenging, as existing teleoperation methods struggle with precision and stability on complex multi-fingered systems. We address this by co-optimizing hand design, perception integration, and teleoperation interface through a universal actuation scheme, custom perception electronics, and two retargeting constraints. We evaluate the platform’s hardware, perception, and teleoperation interface. Training a diffusion policy on collected data shows superior performance over prior works, validating the system’s capability for reliable, high-quality data collection. The platform is constructed from low-cost and off-the-shelf components and will be made public to ensure reproducibility and ease of adoption.

We propose HoliGS, a novel deformable Gaussian splatting framework that addresses embodied view synthesis from long monocular RGB videos. Unlike prior 4D Gaussian splatting and dynamic NeRF pipelines, which struggle with training overhead in minute-long captures, our method leverages invertible Gaussian Splatting deformation networks to reconstruct large-scale, dynamic environments accurately. Specifically, we decompose each scene into a static background plus time-varying objects, each represented by learned Gaussian primitives undergoing global rigid transformations, skeleton-driven articulation, and subtle non-rigid deformations via an invertible neural flow. This hierarchical warping strategy enables robust free-viewpoint novel-view rendering from various embodied camera trajectories by attaching Gaussians to a complete canonical foreground shape (e.g., egocentric or third-person follow), which may involve substantial viewpoint changes and interactions between multiple actors. Our experiments demonstrate that HoliGS achieves superior reconstruction quality on challenging datasets while significantly reducing both training and rendering time compared to state-of-the-art monocular deformable NeRFs. These results highlight a practical and scalable solution for EVS in real-world scenarios. The source code will be released.

Recent advances in multi-modal generative models have enabled significant progress in instruction-based image editing. However, while these models produce visually plausible outputs, their capacity for knowledge-based reasoning editing tasks remains under-explored. In this paper, We introduce KRIS-Bench (Knowledge-based Reasoning in Image-editing Systems Benchmark), a diagnostic benchmark designed to assess models through a cognitively informed lens. Drawing from educational theory, KRIS-Bench categorizes editing tasks across three foundational knowledge types: Factual, Conceptual, and Procedural. Based on this taxonomy, we design 22 representative tasks spanning 7 reasoning dimensions and release 1,267 high-quality annotated editing instances. To support fine-grained evaluation, we propose a comprehensive protocol that incorporates a novel Knowledge Plausibility metric, enhanced by knowledge hints and calibrated through human studies. Empirical results on nine state-of-the-art models reveal significant gaps in reasoning performance, highlighting the need for knowledge-centric benchmarks to advance the development of intelligent image editing systems.

Open-world perception aims to develop a model adaptable to novel domains and various sensor configurations and can understand uncommon objects and corner cases. However, current research lacks sufficiently comprehensive open-world 3D perception benchmarks and robust generalizable methodologies. This paper introduces OpenAD, the first real open-world autonomous driving benchmark for 3D object detection. OpenAD is built upon a corner case discovery and annotation pipeline that integrates with a multimodal large language model (MLLM). The proposed pipeline annotates corner case objects in a unified format for five autonomous driving perception datasets with 2000 scenarios. In addition, we devise evaluation methodologies and evaluate various open-world and specialized 2D and 3D models. Moreover, we propose a vision-centric 3D open-world object detection baseline and further introduce an ensemble method by fusing general and specialized models to address the issue of lower precision in existing open-world methods for the OpenAD benchmark. We host an online challenge on EvalAI. Data, toolkit codes, and evaluation codes are available at https://github.com/VDIGPKU/OpenAD.

Recent advances in 3D scene reconstruction enable real-time viewing in virtual and augmented reality. To support interactive operations for better immersiveness, such as moving or editing objects, 3D scene inpainting methods are proposed to repair or complete the altered geometry. To support users in interacting (such as moving or editing objects) with the scene for the next level of immersiveness, 3D scene inpainting methods are developed to repair the altered geometry. However, current approaches rely on lengthy and computationally intensive optimization, making them impractical for real-time or online applications. We propose InstaInpaint, a reference-based feed-forward framework that produces 3D-scene inpainting from a 2D inpainting proposal within 0.4 seconds. We develop a self-supervised masked-finetuning strategy to enable training of our custom large reconstruction model (LRM) on the large-scale dataset. Through extensive experiments, we analyze and identify several key designs that improve generalization, textural consistency, and geometric correctness. InstaInpaint achieves a 1000$\times$ speed-up from prior methods while maintaining a state-of-the-art performance across two standard benchmarks. Moreover, we show that InstaInpaint generalizes well to flexible downstream applications such as object insertion and multi-region inpainting.

Current 3D scene understanding methods are limited by offline-collected multi-view data or pre-constructed 3D geometry. In this paper, we present ExtractAnything3D (EA3D), a unified online framework for open-world 3D object extraction that enables simultaneous geometric reconstruction and holistic scene understanding. Given a streaming video, EA3D dynamically interprets each frame using vision-language and 2D vision foundation encoders to extract object-level knowledge. This knowledge is integrated and embedded into a Gaussian feature map via a feed-forward online update strategy. We then iteratively estimate visual odometry from historical frames and incrementally update online Gaussian features with new observations. A recurrent joint optimization module directs the model's attention to regions of interest, simultaneously enhancing both geometric reconstruction and semantic understanding. Extensive experiments across diverse benchmarks and tasks, including photo-realistic rendering, semantic and instance segmentation, 3D bounding box and semantic occupancy estimation, and 3D mesh generation, demonstrate the effectiveness of EA3D. Our method establishes a unified and efficient framework for joint online 3D reconstruction and holistic scene understanding, enabling a broad range of downstream tasks. The project webpage is available at \url{https://github.com/VDIGPKU/EA3D}.

We present 4KAgent, a unified agentic super-resolution generalist system designed to universally upscale any image to 4K resolution (and even higher, if applied iteratively). Our system can transform images from extremely low resolutions with severe degradations, for example, highly distorted inputs at $256\times 256$, into crystal-clear, photorealistic 4K outputs. 4KAgent comprises three core components: (1) Profiling, a module that customizes the 4KAgent pipeline based on bespoke use cases; (2) A Perception Agent, which leverages vision-language models alongside image quality assessment experts to analyze the input image and make a tailored restoration plan; and (3) A Restoration Agent, which executes the plan, following a recursive execution-reflection paradigm, guided by a quality-driven mixture-of-experts policy to select the optimal output for each step. Additionally, 4KAgent embeds a specialized face restoration pipeline, significantly enhancing facial details in portrait and selfie photos. We rigorously evaluate our 4KAgent across 11 distinct task categories encompassing a total of 26 diverse benchmarks, setting new state-of-the-art on a broad spectrum of imaging domains. Our evaluations cover natural images, portrait photos, AI-generated content, satellite imagery, fluorescence microscopy, and medical imaging like fundoscopy, ultrasound, and X-ray, demonstrating superior performance in terms of both perceptual (e.g., NIQE, MUSIQ) and fidelity (e.g., PSNR) metrics. By establishing a novel agentic paradigm for low-level vision tasks, we aim to catalyze broader interest and innovation within vision-centric autonomous agents across diverse research communities. We release all the code, models, and results at: https://4kagent.github.io.

Articulated 3D reconstruction has valuable applications in various domains, yet it remains costly and demands intensive work from domain experts. Recent advancements in template-free learning methods show promising results with monocular videos. Nevertheless, these approaches necessitate a comprehensive coverage of all viewpoints of the subject in the input video, thus limiting their applicability to casually captured videos from online sources. In this work, we study articulated 3D shape reconstruction from a single and casually captured Internet video, where the subject's view coverage is incomplete. We propose DreaMo that jointly performs shape reconstruction while solving the challenging low-coverage regions with view-conditioned diffusion prior and several tailored regularizations. In addition, we introduce a skeleton generation strategy to create human-interpretable skeletons from the learned neural bones and skinning weights without any predefined skeleton structures. We conduct our study on a self-collected internet video collection characterized by incomplete view coverage. DreaMo shows promising quality in novel-view rendering, detailed articulated shape reconstruction, and skeleton generation. Extensive qualitative and quantitative studies validate the efficacy of each proposed component, and show existing methods are unable to solve correct geometry due to the incomplete view coverage.

We tackle the challenge of generating long-take videos encompassing multiple non-repetitive yet coherent events. Existing approaches generate long videos conditioned on single input guidance, often leading to repetitive content. To address this problem, we develop a framework that uses multiple guidance sources to enhance long video generation. The main idea of our approach is to decouple video generation into keyframe generation and frame interpolation. In this process, keyframe generation focuses on cre-ating multiple coherent events, while the frame interpolation stage generates smooth intermediate frames between keyframes using existing video generation models. A novel mask attention module is further introduced to improve co-herence and efficiency. Experiments on challenging real-world videos demonstrate that the proposed method outper-forms prior methods by up to 9.5% in objective metrics.

While text-to-video generation shows state-of-the-art results, fine-grained output control remains challenging for users relying solely on natural language prompts. In this work, we present FACTOR for fine-grained controllable video generation. FACTOR provides an intuitive interface where users can manipulate the trajectory and appearance of individual objects in conjunction with a text prompt. We propose a unified framework to integrate these control signals into an existing text-to-video model. Our approach involves a multimodal condition module with a joint encoder, control-attention layers, and an appearance augmentation mechanism. This design enables FACTOR to generate videos that closely align with detailed user specifications. Extensive experiments on standard benchmarks and user-provided inputs demonstrate a notable improvement in controllability by FACTOR over competitive baselines.

Modeling a human avatar that can plausibly deform to articulations is an active area of research. We present PocoLoco - the first template-free, point-based, pose-conditioned generative model for 3D humans in loose clothing. We motivate our work by noting that most methods require a parametric model of the human body to ground pose-dependent deformations. Consequently, they are restricted to modeling clothing that is topologically similar to the naked body and do not extend well to loose clothing. The few methods that attempt to model loose clothing typically require either canonicalization or a UV-parameterization and need to address the challenging problem of explicitly estimating correspondences for the deforming clothes. In this work, we formulate avatar clothing deformation as a conditional point-cloud generation task within the denoising diffusion framework. Crucially, our framework operates directly on unordered point clouds, eliminating the need for a parametric model or a clothing template. This also enables a variety of practical applications, such as point-cloud completion and pose-based editing - important features for virtual human animation. As current datasets for human avatars in loose clothing are far too small for training diffusion models, we release a dataset of two subjects performing various poses in loose clothing with a total of 75K point clouds. By contributing towards tackling the challenging task of effectively modeling loose clothing and expanding the available data for training these models, we aim to set the stage for further innovation in digital humans. The source code is available at https://github.com/sidsunny/pocoloco.

Recently, transformer-based approaches have shown promising results for semi-supervised video object segmentation. However, these approaches typically struggle on long videos due to increased GPU memory demands, as they frequently expand the memory bank every few frames. We propose a transformer-based approach, named MAVOS, that introduces an optimized and dynamic long-term modulated cross-attention (MCA) memory to model temporal smoothness without requiring frequent memory expansion. The proposed MCA effectively encodes both local and global features at various levels of granularity while efficiently maintaining consistent speed regardless of the video length. Extensive experiments on multiple benchmarks, LVOS, Long-Time Video, and DAVIS 2017, demonstrate the effectiveness of our proposed contributions leading to real-time inference and markedly reduced memory demands without any degradation in segmentation accuracy on long videos. Compared to the best existing transformer-based approach, our MAVOS increases the speed by 7.6x, while significantly reducing the GPU memory by 87% with comparable segmentation performance on short and long video datasets. Notably, on the LVOS dataset, our MAVOS achieves a J&F score of 63.3% while operating at 37 frames per second (FPS) on a single V100 GPU. Our code and models will be publicly released at https://github.com/Amshaker/MAVOS.

We present a first step towards 4D (3D space and time) human video stylization, which addresses style transfer, novel view synthesis, and human animation within a unified framework. While numerous video stylization methods have been developed, they are typically restricted to rendering images in specific viewpoints of the input video, lacking the capability to generalize to novel views and novel poses in dynamic scenes. To overcome these limitations, we leverage Neural Radiance Fields (NeRFs) to represent and stylize videos within a single framework. Our method involves simultaneously representing the human subject and the surrounding scene using two NeRFs. This dual representation facilitates the animation of human subjects across various poses and novel viewpoints. A key innovation is our introduction of a geometry-guided tri-plane representation, which significantly boosts the efficiency and robustness of the feature representation compared to direct tri-plane optimization. Stylization is performed within the NeRF rendered feature space, which can reduce the computational burden compared to applying style transformation to the feature vector of sampled points. Extensive experiments demonstrate that the proposed method strikes a superior balance between stylized textures and temporal coherence, surpassing existing approaches. Furthermore, our framework uniquely extends its capabilities to accommodate novel poses and viewpoints, making it a versatile tool for creative human video stylization. The source code and results will be available at this github site . The stylized videos are available in this Youtube video . • We present a 3D video stylization framework for novel views and human poses. • A geometric prior is introduced to enhance tri-plane feature learning efficiency. • Our method balances stylized textures and temporal coherence for dynamic scenes.

Vision systems that see and reason about the compositional nature of visual scenes are fundamental to understanding our world. The complex relations between objects and their locations, ambiguities, and variations in the real-world environment can be better described in human language, naturally governed by grammatical rules and other modalities such as audio and depth. The models learned to bridge the gap between such modalities and large-scale training data facilitate contextual reasoning, generalization, and prompt capabilities at test time. These models are referred to as foundation models. The output of such models can be modified through human-provided prompts without retraining, e.g., segmenting a particular object by providing a bounding box, having interactive dialogues by asking questions about an image or video scene or manipulating the robot's behavior through language instructions. In this survey, we provide a comprehensive review of such emerging foundation models, including typical architecture designs to combine different modalities (vision, text, audio, etc.), training objectives (contrastive, generative), pre-training datasets, fine-tuning mechanisms, and the common prompting patterns; textual, visual, and heterogeneous. We discuss the open challenges and research directions for foundation models in computer vision, including difficulties in their evaluations and benchmarking, gaps in their real-world understanding, limitations of contextual understanding, biases, vulnerability to adversarial attacks, and interpretability issues. We review recent developments in this field, covering a wide range of applications of foundation models systematically and comprehensively.

In this paper, we propose a novel translation model, UniTranslator, for transforming representations between visually distinct domains under conditions of limited training data and significant visual differences. The main idea behind our approach is leveraging the domain-neutral capabilities of CLIP as a bridging mechanism, while utilizing a separate module to extract abstract, domain-agnostic semantics from the embeddings of both the source and target realms. Fusing these abstract semantics with target-specific semantics results in a transformed embedding within the CLIP space. To bridge the gap between the disparate worlds of CLIP and StyleGAN, we introduce a new non-linear mapper, the CLIP2P mapper. Utilizing CLIP embeddings, this module is tailored to approximate the latent distribution in the StyleGAN's latent space, effectively acting as a connector between these two spaces. The proposed UniTranslator is versatile and capable of performing various tasks, including style mixing, stylization, and translations, even in visually challenging scenarios across different visual domains. Notably, UniTranslator generates high-quality translations that showcase domain relevance, diversity, and improved image quality. UniTranslator surpasses the performance of existing general-purpose models and performs well against specialized models in representative tasks.

Burst Image Restoration aims to reconstruct a high-quality image by efficiently combining complementary inter-frame information. However, it is quite challenging since individual burst images often have inter-frame misalignments that usually lead to ghosting and zipper artifacts. To mitigate this, we develop a novel approach for burst image processing named BIPNet that focuses solely on the information exchange between burst frames and filter-out the inherent degradations while preserving and enhancing the actual scene details. Our central idea is to generate a set of pseudo-burst features that combine complementary information from all the burst frames to exchange information seamlessly. However, due to inter-frame misalignment, the information cannot be effectively combined in pseudo-burst. Thus, we initially align the incoming burst features regarding the reference frame using the proposed edge-boosting feature alignment. Lastly, we progressively upscale the pseudo-burst features in multiple stages while adaptively combining the complementary information. Unlike the existing works, that usually deploy single-stage up-sampling with a late fusion scheme, we first deploy a pseudo-burst mechanism followed by the adaptive-progressive feature up-sampling. The proposed BIPNet significantly outperforms the existing methods on burst super-resolution, low-light image enhancement, low-light image super-resolution, and denoising tasks.

Given a piece of text, a video clip, and reference audio, the movie dubbing (also known as Visual Voice Cloning, V2C) task aims to generate speeches that clone reference voice and align well with the video in both emotion and lip movement, which is more challenging than conventional text-to-speech synthesis tasks. To align the generated speech with the inherent lip motion of the given silent video, most existing works utilize each video frame to query textual phonemes. However, such an attention operation usually leads to mumble speech because different phonemes are fused for video frames corresponding to one phoneme (video frames are finer-grained than phonemes). To address this issue, we propose a diffusion-based movie dubbing architecture, which improves pronunciation by Hierarchical Phoneme Modeling (HPM) and generates better mel-spectrogram through Acoustic Diffusion Denoising (ADD). We term our model as HD-Dubber. Specifically, our HPM bridges the visual information and corresponding speech prosody from three aspects: (1) aligning lip movement with the speech duration based on each phoneme unit by contrastive learning; (2) conveying facial expression to phoneme-level energy and pitch; and (3) injecting global emotions captured from video scenes into prosody. On the other hand, ADD exploits a denoising diffusion framework to transform the noise signal into a mel-spectrogram via a parameterized Markov chain conditioned on textual phonemes and reference audio. ADD has two novel denoisers, the Style-adaptive Residual Denoiser (SRD) and the Phoneme-enhanced U-net Denoiser (PUD), to enhance speaker similarity and improve pronunciation quality. Extensive experimental results on the three benchmark datasets demonstrate the state-of-the-art performance of the proposed method. The source code and trained models will be made available to the public.

Deep learning methods have demonstrated state-of-the-art performance in image restoration, especially when trained on large-scale paired datasets. However, acquiring paired data in real-world scenarios poses a significant challenge. Unsupervised restoration approaches based on generative adversarial networks (GANs) offer a promising solution without requiring paired datasets. Yet, these GAN-based approaches struggle to surpass the performance of conventional unsupervised GAN-based frameworks without significantly modifying model structures or increasing the computational complexity. To address these issues, we propose a self-collaboration (SC) strategy for existing restoration models. This strategy utilizes information from the previous stage as feedback to guide subsequent stages, achieving significant performance improvement without increasing the framework's inference complexity. The SC strategy comprises a prompt learning (PL) module and a restorer ($Res$Res). It iteratively replaces the previous less powerful fixed restorer $\overline{Res}$Res¯ in the PL module with a more powerful $Res$Res. The enhanced PL module generates better pseudo-degraded/clean image pairs, leading to a more powerful $Res$Res for the next iteration. Our SC can significantly improve the $Res$Res 's performance by over 1.5 dB without adding extra parameters or computational complexity during inference. Meanwhile, existing self-ensemble (SE) and our SC strategies enhance the performance of pre-trained restorers from different perspectives. As SE increases computational complexity during inference, we propose a re-boosting module to the SC (Reb-SC) to improve the SC strategy further by incorporating SE into SC without increasing inference time. This approach further enhances the restorer's performance by approximately 0.3 dB. Additionally, we present a baseline framework that includes parallel generative adversarial branches with complementary "self-synthesis" and "unpaired-synthesis" constraints, ensuring the effectiveness of the training framework. Extensive experimental results on restoration tasks demonstrate that the proposed model performs favorably against existing state-of-the-art unsupervised restoration methods.

In this paper, we study the application of Vehicle-to-Everything (V2X) communication to improve the perception performance of autonomous vehicles. We present V2X-ViTs, a robust cooperative perception framework with V2X communication using novel vision Transformer models. First, we present V2X-ViTv1 containing holistic attention modules that can effectively fuse information across on-road agents (i.e., vehicles and infrastructure). Specifically, V2X-ViTv1 consists of alternating layers of heterogeneous multi-agent self-attention and multi-scale window self-attention, which captures inter-agent interaction and per-agent spatial relationships. These key modules are designed in a unified Transformer architecture to handle common V2X challenges, including asynchronous information sharing, pose errors, and heterogeneity of V2X components. Second, we propose an advanced architecture, V2X-ViTv2, that enjoys increased ability for multi-scale perception. We also propose advanced data augmentation techniques tailored for V2X applications to improve performance. We construct a large-scale V2X perception dataset using CARLA and OpenCDA to validate our approach. Extensive experimental results on both synthetic and real-world datasets show that V2X-ViTs achieve state-of-the-art performance for 3D object detection and are robust even under harsh, noisy environments.

This paper highlights a problem of evaluation metrics adopted in the open-vocabulary segmentation. The evaluation process relies heavily on closed-set metrics on zero-shot or cross-dataset pipelines without considering the similarity between predicted and ground truth categories. We first survey eleven similarity measurements between two categorical words using WordNet linguistics statistics, text embedding, or language models by comprehensive quantitative analysis and user study to tackle this issue. Based on those explored measurements, we design novel evaluation metrics, Open mIoU, Open AP, and Open PQ, tailored for three open-vocabulary segmentation tasks. We benchmark the proposed evaluation metrics on twelve open-vocabulary methods in three segmentation tasks. Despite the relative subjectivity of similarity distance, we demonstrate that our metrics can still well evaluate the open ability of the existing open-vocabulary segmentation methods. We hope our work can bring the community new thinking about evaluating model ability for open-vocabulary segmentation.

Generalized zero-shot learning (GZSL) requires that models are able to recognize classes they were trained on, and new classes they haven't seen before. Feature-generation approaches are popular due to their effectiveness in mitigating overfitting to the training classes. Existing generative approaches usually adopt simple discriminators for distribution or classification supervision, however, thus limiting their ability to generate visual features that are discriminative of and transferable to novel categories. To overcome this limitation and improve the quality of generated features, we propose a dual prototype contrastive augmented discriminator for the generative adversarial network. Specifically, we design a Dual Prototype Contrastive Network (DPCN), which leverages complementary information between visual space and semantic space through multi-task prototype contrastive learning. Contrastive learning of the visual prototypes enhances the ability of the generated features to distinguish between classes, while the contrastive learning of the semantic prototypes improves their transferability. Furthermore, we introduce margins into the contrastive learning process to ensure both intra-class compactness and inter-class separation. To demonstrate the effectiveness of the proposed approach, we conduct experiments on three widely-used zero-shot learning benchmark datasets, where DPCN achieves state-of-the-art performance for GZSL.

Most existing non-blind deblurring methods formulate the problem into a maximum-a-posteriori framework and address it by manually designing a variety of regularization terms and data terms of the latent clear images. However, explicitly designing these two terms is quite challenging, which usually leads to complex optimization problems. In this paper, we propose a Discriminative Shrinkage Deep Network for fast and accurate deblurring. Most existing methods use deep convolutional neural networks (CNNs), or radial basis functions only to learn the regularization term. In contrast, we formulate both the data and regularization terms while splitting the deconvolution model into data-related and regularization-related sub-problems. We explore the properties of the Maxout function and develop a deep CNN model with Maxout layers to learn discriminative shrinkage functions, which directly approximate the solutions of these two sub-problems. Moreover, we develop a U-Net according to Krylov subspace method to restore the latent clear images effectively and efficiently, which plays a role but is better than the conventional fast-Fourier-transform-based or conjugate gradient method. Experimental results show that the proposed method performs favorably against the state-of-the-art methods regarding efficiency and accuracy.

Rain in the dark poses significant challenges to deploying real-world applications such as autonomous driving, surveillance systems, and night photography. Existing low-light enhancement or deraining methods struggle to brighten low-light conditions and remove rain simultaneously. Cascade approaches, like “deraining followed by low-light enhancement” and vice versa, often result in problematic rain patterns or overly blurred and overexposed images. To address these challenges, we introduce a novel two-stage model called L2RIRNet, which innovatively integrates low-light enhancement and deraining into a single framework in real-world settings. Our model comprises two key components: a Dual Degradation Representation Network (DDR-Net) and a Restoration Network. The DDR-Net independently learns degradation representations for luminance effects in dark areas and rain patterns in light areas, which are constrained by dual degradation loss and have not been discussed in the previous methods. The Restoration Network restores the degraded image using a Fourier Detail Guidance (FDG) module, which focuses on texture details in frequency and spatial domains to inform the restoration process and leverages near-rainless detailed images. Furthermore, we contribute a dataset containing both synthetic and real-world low-light-rainy images. Extensive experiments demonstrate that our L2RIRNet performs favorably against existing methods in synthetic and complex real-world scenarios. All the code and dataset can be found inhttps://github.com/linxin0/Low_light_rainy.

Recently, researchers have proposed many deep generative models, including generative adversarial networks (GANs) and denoising diffusion models. Although significant breakthroughs have been made and empirical success has been achieved with the GAN, its mathematical underpinnings remain relatively unknown. This paper focuses on a rigorous mathematical theoretical framework: the composite-functional-gradient GAN (CFG). Specifically, we reveal the theoretical connection between the CFG model and score-based models. We find that the CFG discriminator’s training objective is equivalent to finding an optimal$D(\mathrm {x})$. The optimal$D(\mathrm {x})$’s gradient differentiates the integral of the differences between the score functions of real and synthesized samples. Conversely, training the CFG generator involves finding an optimal$G(\mathrm {x})$that minimizes this difference. In this paper, we aim to derive an annealed weight preceding the CFG discriminator’s weight. This new explicit theoretical explanation model is called the annealed CFG method. To overcome the annealed CFG method’s limitation, as the method is not readily applicable to the state-of-the-art (SOTA) GAN model, we propose a nested annealed training scheme (NATS). This scheme keeps the annealed weight from the CFG method and can be seamlessly adapted to various GAN models, no matter their structural, loss, or regularization differences. We conduct thorough experimental evaluations on various benchmark datasets for image generation. The results show that our annealed CFG and NATS methods significantly improve the synthesized samples’ quality and diversity. This improvement is clear when comparing the CFG method and the SOTA GAN models.

Due to the limited output categories, semi-supervised salient object detection faces challenges in adapting conventional semi-supervised strategies. To address this limitation, we propose a multi-branch architecture that extracts complementary features from labeled data. Specifically, we introduce TripleNet, a three-branch network architecture designed for contour, content, and holistic saliency prediction. The supervision signals for the contour and content branches are derived by decomposing the limited ground truths. After training on the labeled data, the model produces pseudo-labels for unlabeled images, including contour, content, and salient objects. By leveraging the complementarity between the contour and content branches, we construct coupled pseudo-saliency labels by integrating the pseudo-contour and pseudo-content labels, which differ from the model-inferred pseudo-saliency labels. We further develop an enhanced pseudo-labeling mechanism that generates enhanced pseudo-saliency labels by combining reliable regions from both pseudo-saliency labels. Moreover, we incorporate a partial binary cross-entropy loss function to guide the learning of the saliency branch to focus on effective regions within the enhanced pseudo-saliency labels, which are identified through our adaptive thresholding approach. Extensive experiments demonstrate that the proposed method achieves state-of-the-art performance using only 329 labeled training images.

The weakly supervised video anomaly detection aims to learn a detection model using only video-level labeled data. Prior studies ignore the complexity or duration of anomalies present in abnormal videos during temporal modeling. Moreover, existing works usually detect the most abnormal segments, potentially overlooking the completeness of anomalies. We propose a dynamic erasing network (DE-Net) for weakly supervised video anomaly detection, which learns video-specific temporal features via adaptive temporal modeling (ATM) to address these limitations. Specifically, to handle duration variations of abnormal events, we propose an ATM module capable of adaptively selecting and aggregating the most appropriate K temporal scale features for each video. Then, we design a dynamic erasing (DE) strategy that dynamically assesses the completeness of the detected anomalies and erases prominent abnormal segments to encourage the model to discover gentle abnormal segments. The proposed method achieves favorable performance compared to several state-of-the-art approaches on the widely used XD-Violence, TAD, and UCF-Crime datasets.

We present a physics-based inverse rendering method that learns the illumination, geometry, and materials of a scene from posed multi-view RGB images. To model the illumination of a scene, existing inverse rendering works either completely ignore the indirect illumination or model it by coarse approximations, leading to sub-optimal illumination, geometry, and material prediction of the scene. In this work, we propose a physics-based illumination model that first locates surface points through an efficient refined sphere tracing algorithm, then explicitly traces the incoming indirect lights at each surface point based on reflection. Then, we estimate each identified indirect light through an efficient neural network. Moreover, we utilize the Leibniz’s integral rule to resolve non-differentiability in the proposed illumination model caused by boundary lights inspired by differentiable irradiance in computer graphics. As a result, the proposed differentiable illumination model can be learned end-to-end together with geometry and materials estimation. As a side product, our physics-based inverse rendering model also facilitates flexible and realistic material editing as well as relighting. Extensive experiments on synthetic and real-world datasets demonstrate that the proposed method performs favorably against existing inverse rendering methods on novel view synthesis and inverse rendering. The source code and trained models can be found at the project page: https://denghilbert.github.io/pii.

Existing LiDAR-based 3D object detection methods for autonomous driving scenarios mainly adopt the training-from-scratch paradigm. Unfortunately, this paradigm heavily relies on large-scale labeled data, whose collection can be expensive and time-consuming. Self-supervised pre-training is an effective and desirable way to alleviate this dependence on extensive annotated data. In this work, we present BEV-MAE, an efficient masked autoencoder pre-training framework for LiDAR-based 3D object detection in autonomous driving. Specifically, we propose a bird's eye view (BEV) guided masking strategy to guide the 3D encoder learning feature representation in a BEV perspective and avoid complex decoder design during pre-training. Furthermore, we introduce a learnable point token to maintain a consistent receptive field size of the 3D encoder with fine-tuning for masked point cloud inputs. Based on the property of outdoor point clouds in autonomous driving scenarios, i.e., the point clouds of distant objects are more sparse, we propose point density prediction to enable the 3D encoder to learn location information, which is essential for object detection. Experimental results show that BEV-MAE surpasses prior state-of-the-art self-supervised methods and achieves a favorably pre-training efficiency. Furthermore, based on TransFusion-L, BEV-MAE achieves new state-of-the-art LiDAR-based 3D object detection results, with 73.6 NDS and 69.6 mAP on the nuScenes benchmark. The source code will be released at https://github.com/VDIGPKU/BEV-MAE.

Addressing Out-Of-Distribution (OOD) Segmentation and Zero-Shot Semantic Segmentation (ZS3) is challenging, necessitating segmenting unseen classes. Existing strategies adapt the class-agnostic Mask2Former (CA-M2F) tailored to specific tasks. However, these methods cater to singular tasks, demand training from scratch, and we demonstrate certain deficiencies in CA-M2F, which affect performance. We propose the Class-Agnostic Structure-Constrained Learning (CSL), a plug-in framework that can integrate with existing methods, thereby embedding structural constraints and achieving performance gain, including the unseen, specifically OOD, ZS3, and domain adaptation (DA) tasks. There are two schemes for CSL to integrate with existing methods (1) by distilling knowledge from a base teacher network, enforcing constraints across training and inference phrases, or (2) by leveraging established models to obtain per-pixel distributions without retraining, appending constraints during the inference phase. Our soft assignment and mask split methodologies enhance OOD object segmentation. Empirical evaluations demonstrate CSL's prowess in boosting the performance of existing algorithms spanning OOD segmentation, ZS3, and DA segmentation, consistently transcending the state-of-art across all three tasks.

Contents generated by recent advanced Text-to-Image (T2I) diffusion models are sometimes too imaginative for existing off-the-shelf dense predictors to estimate due to the immitigable domain gap. We introduce DMP, a pipeline utilizing pre-trained T2I models as a prior for dense prediction tasks. To address the misalignment between deterministic prediction tasks and stochastic T2I models, we reformulate the diffusion process through a sequence of interpo-lations, establishing a deterministic mapping between input RGB images and output prediction distributions. To preserve generalizability, we use low-rank adaptation to fine-tune pre-trained models. Extensive experiments across five tasks, including 3D property estimation, semantic segmentation, and intrinsic image decomposition, showcase the efficacy of the proposed method. Despite limited-domain training data, the approach yields faithful estimations for arbitrary images, surpassing existing state-of-the-art algorithms. The code is available at https://github.com/shinying/dmp.

While pre-trained large-scale vision models have shown significant promise for semantic correspondence, their features often struggle to grasp the geometry and orientation of instances. This paper identifies the importance of being geometry-aware for semantic correspondence and reveals a limitation of the features of current foundation models under simple post-processing. We show that incorporating this information can markedly enhance semantic correspondence performance with simple but effective solutions in both zero-shot and supervised settings. We also construct a new challenging benchmark for semantic correspondence built from an existing animal pose estimation dataset, for both pre-training validating models. Our method achieves a [email protected] score of 65.4 (zero-shot) and 85.6 (supervised) on the challenging SPair-71k dataset, surpassing the state of the art by 5.5p and 11.0p absolute gains, respectively. Our code and datasets are publicly available at: https://telling-left-from-right.github.io.

In subject-driven text-to-image synthesis, the synthesis process tends to be heavily influenced by the reference images provided by users, often overlooking crucial attributes detailed in the text prompt. In this work, we propose Subject-Agnostic Guidance (SAG), a simple yet effective solution to remedy the problem. We show that through constructing a subject-agnostic condition and applying our proposed dual classifier-free guidance, one could obtain outputs consistent with both the given subject and input text prompts. We validate the efficacy of our approach through both optimization-based and encoder-based methods. Additionally, we demonstrate its applicability in second-order customization methods, where an encoder-based model is fine-tuned with DreamBooth. Our approach is conceptually simple and requires only minimal code modifications, but leads to substantial quality improvements, as evidenced by our evaluations and user studies.

The quality of the data and annotation upper-bounds the quality of a downstream model. While there exist large text corpora and image-text pairs, high-quality video-text data is much harder to collect. First of all, manual labeling is more time-consuming, as it requires an annotator to watch an entire video. Second, videos have a temporal dimension, consisting of several scenes stacked together, and showing multiple actions. Accordingly, to establish a video dataset with high-quality captions, we propose an automatic approach leveraging multimodal inputs, such as textual video description, subtitles, and individual video frames. Specifically, we curate 3.8M high-resolution videos from the publicly available HD-VILA-100M dataset. We then split them into semantically consistent video clips, and apply multiple cross-modality teacher models to obtain captions for each video. Next, we finetune a retrieval model on a small subset where the best caption of each video is manually selected and then employ the model in the whole dataset to select the best caption as the annotation. In this way, we get 70M videos paired with high-quality text captions. We dub the dataset as Panda-70M. We show the value of the proposed dataset on three downstream tasks: video captioning, video and text retrieval, and text-driven video generation. The models trained on the proposed data score substantially better on the majority of metrics across all the tasks.

Recent temporal LiDAR-based 3D object detectors achieve promising performance based on the two-stage proposal-based approach. They generate 3D box candidates from the first-stage dense detector, followed by different temporal aggregation methods. However, these approaches require per-frame objects or whole point clouds, posing challenges related to memory bank utilization. Moreover, point clouds and trajectory features are combined solely based on concatenation, which may neglect effective interactions between them. In this paper, we propose a point-trajectory transformer with long short-term memory for efficient temporal 3D object detection. To this end, we only utilize point clouds of current-frame objects and their historical trajectories as input to minimize the memory bank storage requirement. Furthermore, we introduce modules to encode trajectory features, focusing on long short-term and future-aware perspectives, and then effectively aggregate them with point cloud features. We conduct extensive experiments on the large-scale Waymo dataset to demon-strate that our approach performs well against state-of-the-art methods. Code and models will be made publicly available at https://github.com/kuanchihhuang/Ptt.

Existing tracking methods mainly focus on learning better target representation or developing more robust prediction models to improve tracking performance. While tracking performance has significantly improved, the target loss issue occurs frequently due to tracking failures, complete occlusion, or out-of-view situations. However, con-siderably less attention is paid to the self-recovery issue of tracking methods, which is crucial for practical applications. To this end, we propose a recoverable tracking framework, RTracker, that uses a tree-structured memory to dynamically associate a tracker and a detector to enable self-recovery ability. Specifically, we propose a Positive-Negative Tree-structured memory to chronologically store and maintain positive and negative target samples. Upon the PN tree memory, we develop corresponding walking rules for determining the state of the target and define a set of control flows to unite the tracker and the detector in different tracking scenarios. Our core idea is to use the support samples of positive and negative target categories to establish a relative distance-based criterion for a reliable assessment of target loss. The favorable performance in comparison against the state-of-the-art methods on nu-merous challenging benchmarks demonstrates the effectiveness of the proposed algorithm. All the source code and trained models will be released at https://github.com/NorahGreen/RTracker.

Diffusion models have made significant advances in generating high-quality images, but their application to video generation has remained challenging due to the complexity of temporal motion. Zero-shot video editing offers a solution by utilizing pre-trained image diffusion models to translate source videos into new ones. Nevertheless, existing methods struggle to maintain strict temporal consistency and efficient memory consumption. In this work, we propose a novel approach to enhance temporal consistency in generated videos by merging self-attention tokens across frames. By aligning and compressing temporally redundant tokens across frames, our method improves temporal coherence and reduces memory consumption in self-attention computations. The merging strategy matches and aligns tokens according to the temporal correspondence between frames, facilitating natural temporal consistency in generated video frames. To manage the complexity of video processing, we divide videos into chunks and develop intra-chunk local token merging and inter-chunk global token merging, ensuring both short-term video continuity and long-term content consistency. Our video editing approach seamlessly extends the advancements in image editing to video editing, rendering favorable results in temporal consistency over state-of-the-art methods.

Language has emerged as a natural interface for image editing. In this paper, we introduce a method for region- based image editing driven by textual prompts, without the need for user-provided masks or sketches. Specifically, our approach leverages an existing pre-trained text-to-image model and introduces a bounding box generator to iden-tify the editing regions that are aligned with the textual prompts. We show that this simple approach enables flex-ible editing that is compatible with current image generation models, and is able to handle complex prompts featuring multiple objects, complex sentences, or lengthy para- graphs. We conduct an extensive user study to compare our method against state-of-the-art methods. The experiments demonstrate the competitive performance of our method in manipulating images with high fidelity and realism that correspond to the provided language descriptions. Our project webpage can be found at: https://yuanzelin.me/LearnableRegions_page.

Video Individual Counting (VIC) aims to predict the number of unique individuals in a single video. Existing methods learn representations based on trajectory labels for individuals, which are annotation-expensive. To provide a more realistic reflection of the underlying practical challenge, we introduce a weakly supervised VIC task, wherein trajectory labels are not provided. Instead, two types of labels are provided to indicate traffic entering the field of view (inflow) and leaving the field view (outflow). We also propose the first solution as a baseline that formulates the task as a weakly supervised contrastive learning problem under group-level matching. In doing so, we devise an end-to-end trainable soft contrastive loss to drive the network to distin-guish inflow, outflow, and the remaining. To facilitate future study in this direction, we generate annotations from the existing VIC datasets Sense Crowd and CroHD and also build a new dataset, UAVVIC. Extensive results show that our baseline weakly supervised method outperforms supervised methods, and thus, little information is lost in the transition to the more practically relevant weakly supervised task. The code and trained model can be found at CGNet.

Eliminating image blur produced by various kinds ofmotion has been a challenging problem. Dominant approaches rely heavily on model capacity to remove blurring by reconstructing residual from blurry observation in feature space. These practices not only prevent the capture of spatially variable motion in the real world but also ignore the tai-lored handling of various motions in image space. In this paper, we propose a novel real-world deblurring filtering model called the Motion-adaptive Separable Collaborative (MISC) Filter. In particular, we use a motion estimation net-work to capture motion information from neighborhoods, thereby adaptively estimating spatially-variant motion flow, mask, kernels, weights, and offsets to obtain the MISC Fil-ter. The MISC Filter first aligns the motion-induced blur-ring patterns to the motion middle along the predicted flow direction, and then collaboratively filters the aligned image through the predicted kernels, weights, and offsets to generate the output. This design can handle more general-ized and complex motion in a spatially differentiated man-ner. Furthermore, we analyze the relationships between the motion estimation network and the residual reconstruction network. Extensive experiments on four widely used bench-marks demonstrate that our method provides an effective solution for real-world motion blur removal and achieves state-of-the-art performance. Code is available at https://github.com/ChengxuLiu/MISCFilter.

This paper introduces a novel unified representation of diffusion models for image generation and segmentation. Specifically, we use a colormap to represent entity-level masks, addressing the challenge of varying entity numbers while aligning the representation closely with the image RGB domain. Two novel modules, including the location-aware color palette and progressive dichotomy module, are proposed to support our mask representation. On the one hand, a location-aware palette guarantees the colors' consistency to entities' locations. On the other hand, the progressive dichotomy module can efficiently decode the synthesized colormap to high-quality entity-level masks in a depth-first binary search without knowing the cluster numbers. To tackle the issue of lacking large-scale segmentation training data, we employ an inpainting pipeline and then improve the flexibility of diffusion models across various tasks, including inpainting, image synthesis, referring segmentation, and entity segmentation. Comprehensive experiments validate the efficiency of our approach, demonstrating comparable segmentation mask quality to state-of-the-art and adaptability to multiple tasks.

Large Multimodal Models (LMMs) extend Large Lan-guage Models to the vision domain. Initial LMMs used holistic images and text prompts to generate ungrounded textual responses. Recently, region-level LMMs have been used to generate visually grounded responses. However, they are limited to only referring to a single object category at a time, require users to specify the regions, or can-not offer dense pixel-wise object grounding. In this work, we present Grounding LMM (GLaMM), the first model that can generate natural language responses seamlessly in-tertwined with corresponding object segmentation masks. GLaMM not only grounds objects appearing in the con-versations but is flexible enough to accept both textual and optional visual prompts (region of interest) as input. This empowers users to interact with the model at various levels of granularity, both in textual and visual domains. Due to the lack of standard benchmarks for the novel setting of visually Grounded Conversation Generation (GCG), we in-troduce a comprehensive evaluation protocol with our curated grounded conversations. Our proposed GCG task requires densely grounded concepts in natural scenes at a large-scale. To this end, we propose a densely annotated Grounding-anything Dataset (GranD) using our proposed automated annotation pipeline that encompasses 7.5M unique concepts grounded in a total of 810M regions available with segmentation masks. Besides GCG, GLaMM also performs effectively on several downstream tasks, e.g., referring expression segmentation, image and region-level captioning and vision-language conversations.

Inherent ambiguity in layout annotations poses significant challenges to developing accurate 360° room layout estimation models. To address this issue, we propose a novel Bi-Layout model capable of predicting two distinct layout types. One stops at ambiguous regions, while the other extends to encompass all visible areas. Our model employs two global context embeddings, where each embedding is designed to capture specific contextual information for each layout type. With our novel feature guidance module, the image feature retrieves relevant context from these embeddings, generating layout-aware features for precise bi-layout predictions. A unique property of our Bi-Layout model is its ability to inherently detect ambiguous regions by comparing the two predictions. To circumvent the need for manual correction of ambiguous annotations during testing, we also introduce a new metric for disambiguating ground truth layouts. Our method demonstrates superior performance on benchmark datasets, notably outperforming leading approaches. Specifically, on the MatterportLayout dataset, it improves 3DIoU from 81.70% to 82.57% across the full test set and notably from 54.80% to 59.97% in subsets with significant ambiguity.

Video grounding aims to localize a spatio-temporal section in a video corresponding to an input text query. This paper addresses a critical limitation in current video grounding methodologies by introducing an Open-Vocabulary Spatio- Temporal Video Grounding task. Unlike prevalent closed-set approaches that struggle with open-vocabulary scenarios due to limited training data and pre-defined vocabularies, our model leverages pre-trained rep-resentations from foundational spatial grounding models. This empowers it to effectively bridge the semantic gap be-tween natural language and diverse visual content, achieving strong performance in closed-set and open-vocabulary settings. Our contributions include a novel spatio-temporal video grounding model, surpassing state-of-the-art results in closed-set evaluations on multiple datasets and demon-strating superior performance in open-vocabulary scenar-ios. Notably, the proposed model outperforms state-of-the-art methods in closed-set settings on VidSTG (Declarative and Interrogative) and HC-STVG (VI and V2) datasets. Furthermore, in open-vocabulary evaluations on HC-STVG VI and YouCook-Interactions, our model surpasses the re-cent best-performing models by 4.88 m.vloU and 1.83% ac-curacy, demonstrating its efficacy in handling diverse lin-guistic and visual concepts for improved video understanding. Our codes will be publicly released.

We present DrivingGaussian, an efficient and effective framework for surrounding dynamic autonomous driving scenes. For complex scenes with moving objects, we first sequentially and progressively model the static background of the entire scene with incremental static 3D Gaussians. We then leverage a composite dynamic Gaussian graph to handle multiple moving objects, individually reconstructing each object and restoring their accurate positions and occlusion relationships within the scene. We further use a LiDAR prior for Gaussian Splatting to reconstruct scenes with greater details and maintain panoramic consistency. DrivingGaussian outperforms existing methods in dynamic driving scene reconstruction and enables photorealistic surround-view synthesis with high-fidelity and multi-camera consistency. Our project page is at: https:/github.com/VDIGPKU/DrivingGaussian.

Restoring facial details from low-quality (LQ) images has remained challenging due to the nature of the problem caused by various degradations in the wild. The codebook prior has been proposed to address the ill-posed problems by leveraging an autoencoder and learned codebook of high-quality (HQ) features, achieving remarkable quality. However, existing approaches in this paradigm frequently depend on a single encoder pre-trained on HQ data for restoring HQ images, disregarding the domain gap and distinct feature representations between LQ and HQ images. As a result, encoding LQ inputs with the same encoder could be insufficient, resulting in imprecise feature representation and leading to suboptimal performance. To tackle this problem, we propose a novel dual-branch framework named $\textit{DAEFR}$. Our method introduces an auxiliary LQ branch that extracts domain-specific information from the LQ inputs. Additionally, we incorporate association training to promote effective synergy between the two branches, enhancing code prediction and restoration quality. We evaluate the effectiveness of DAEFR on both synthetic and real-world datasets, demonstrating its superior performance in restoring facial details. Project page: https://liagm.github.io/DAEFR/

Existing video-language pre-training methods primarily focus on instance-level alignment between video clips and captions via global contrastive learning but neglect rich fine-grained local information in both videos and text, which is of importance to downstream tasks requiring temporal localization and semantic reasoning. A powerful model is expected to be capable of capturing region-object correspondences and recognizing scene changes in a video clip, reflecting spatial and temporal granularity, respectively. To strengthen model's understanding into such fine-grained details, we propose a simple yet effective video-language modeling framework, S-ViLM, by exploiting the intrinsic structures of these two modalities. It includes two novel designs, inter-clip spatial grounding and intra-clip temporal grouping, to promote learning region-object alignment and temporal-aware features, simultaneously. Comprehensive evaluations demonstrate that S-ViLM performs favorably against existing approaches in learning more expressive representations. Specifically, S-ViLM surpasses the state-of-the-art methods substantially on four representative downstream tasks, covering text-video retrieval, video question answering, video action recognition, and temporal action localization.

While Large Language Models (LLMs) are the dominant models for generative tasks in language, they do not perform as well as diffusion models on image and video generation. To effectively use LLMs for visual generation, one crucial component is the visual tokenizer that maps pixel-space inputs to discrete tokens appropriate for LLM learning. In this paper, we introduce \modelname{}, a video tokenizer designed to generate concise and expressive tokens for both videos and images using a common token vocabulary. Equipped with this new tokenizer, we show that LLMs outperform diffusion models on standard image and video generation benchmarks including ImageNet and Kinetics. In addition, we demonstrate that our tokenizer surpasses the previously top-performing video tokenizer on two more tasks: (1) video compression comparable to the next-generation video codec (VCC) according to human evaluations, and (2) learning effective representations for action recognition tasks.

We introduce VideoPrism, a general-purpose video encoder that tackles diverse video understanding tasks with a single frozen model. We pretrain VideoPrism on a heterogeneous corpus containing 36M high-quality video-caption pairs and 582M video clips with noisy parallel text (e.g., ASR transcripts). The pretraining approach improves upon masked autoencoding by global-local distillation of semantic video embeddings and a token shuffling scheme, enabling VideoPrism to focus primarily on the video modality while leveraging the invaluable text associated with videos. We extensively test VideoPrism on four broad groups of video understanding tasks, from web video question answering to CV for science, achieving state-of-the-art performance on 31 out of 33 video understanding benchmarks.

We present VideoPoet, a language model capable of synthesizing high-quality video from a large variety of conditioning signals. VideoPoet employs a decoder-only transformer architecture that processes multimodal inputs – including images, videos, text, and audio. The training protocol follows that of Large Language Models (LLMs), consisting of two stages: pretraining and task-specific adaptation. During pretraining, VideoPoet incorporates a mixture of multimodal generative objectives within an autoregressive Transformer framework. The pretrained LLM serves as a foundation that can be adapted for a range of video generation tasks. We present empirical results demonstrating the model’s state-of-the-art capabilities in zero-shot video generation, specifically highlighting the ability to generate high-fidelity motions. Project page: http://sites.research.google/videopoet/

This paper addresses the scarcity of large-scale datasets for accurate object-in-hand pose estimation, which is crucial for robotic in-hand manipulation within the "Perception-Planning-Control" paradigm. Specifically, we introduce VinT-6D, the first extensive multi-modal dataset integrating vision, touch, and proprioception, to enhance robotic manipulation. VinT-6D comprises 2 million VinT-Sim and 0.1 million VinT-Real entries, collected via simulations in Mujoco and Blender and a custom-designed real-world platform. This dataset is tailored for robotic hands, offering models with whole-hand tactile perception and high-quality, well-aligned data. To the best of our knowledge, the VinT-Real is the largest considering the collection difficulties in the real-world environment so it can bridge the gap of simulation to real compared to the previous works. Built upon VinT-6D, we present a benchmark method that shows significant improvements in performance by fusing multi-modal information. The project is available at https://VinT-6D.github.io/.

We present GALA3D, generative 3D GAussians with LAyout-guided control, for effective compositional text-to-3D generation. We first utilize large language models (LLMs) to generate the initial layout and introduce a layout-guided 3D Gaussian representation for 3D content generation with adaptive geometric constraints. We then propose an instance-scene compositional optimization mechanism with conditioned diffusion to collaboratively generate realistic 3D scenes with consistent geometry, texture, scale, and accurate interactions among multiple objects while simultaneously adjusting the coarse layout priors extracted from the LLMs to align with the generated scene. Experiments show that GALA3D is a user-friendly, end-to-end framework for state-of-the-art scene-level 3D content generation and controllable editing while ensuring the high fidelity of object-level entities within the scene. The source codes and models will be available at gala3d.github.io.

Video understanding has witnessed significant progress with recent video foundation models demonstrating strong performance owing to self-supervised pre-training objectives; Masked Autoencoders (MAE) being the design of choice. Nevertheless, the majority of prior works that leverage MAE pre-training have focused on relatively short video representations (16 / 32 frames in length) largely due to hardware memory and compute limitations that scale poorly with video length due to the dense memory-intensive self-attention decoding. One natural strategy to address these challenges is to subsample tokens to reconstruct during decoding (or decoder masking). In this work, we propose an effective strategy for prioritizing tokens which allows training on longer video sequences (128 frames) and gets better performance than, more typical, random and uniform masking strategies. The core of our approach is an adaptive decoder masking strategy that prioritizes the most important tokens and uses quantized tokens as reconstruction objectives. Our adaptive strategy leverages a powerful MAGVIT-based tokenizer that jointly learns the tokens and their priority. We validate our design choices through exhaustive ablations and observe improved performance of the resulting long-video (128 frames) encoders over short-video (32 frames) counterparts. With our long-video masked autoencoder (LVMAE) strategy, we surpass state-of-the-art on Diving48 by 3.9 points and EPIC-Kitchens-100 verb classification by 2.5 points while relying on a simple core architecture and video-only pre-training (unlike some of the prior works that require millions of labeled video-text pairs or specialized encoders).

Image Restoration (IR), a classic low-level vision task, has witnessed significant advancements through deep models that effectively model global information. Notably, the emergence of Vision Transformers (ViTs) has further propelled these advancements. When computing, the self-attention mechanism, a cornerstone of ViTs, tends to encompass all global cues, even those from semantically unrelated objects or regions. This inclusivity introduces computational inefficiencies, particularly noticeable with high input resolution, as it requires processing irrelevant information, thereby impeding efficiency. Additionally, for IR, it is commonly noted that small segments of a degraded image, particularly those closely aligned semantically, provide particularly relevant information to aid in the restoration process, as they contribute essential contextual cues crucial for accurate reconstruction. To address these challenges, we propose boosting IR's performance by sharing the key semantics via Transformer for IR (i.e., SemanIR) in this paper. Specifically, SemanIR initially constructs a sparse yet comprehensive key-semantic dictionary within each transformer stage by establishing essential semantic connections for every degraded patch. Subsequently, this dictionary is shared across all subsequent transformer blocks within the same stage. This strategy optimizes attention calculation within each block by focusing exclusively on semantically related components stored in the key-semantic dictionary. As a result, attention calculation achieves linear computational complexity within each window. Extensive experiments across 6 IR tasks confirm the proposed SemanIR's state-of-the-art performance, quantitatively and qualitatively showcasing advancements. The visual results, code, and trained models are available at: https://github.com/Amazingren/SemanIR.

Semantic segmentation and semantic image synthesis are two representative tasks in visual perception and generation. While existing methods consider them as two distinct tasks, we propose a unified framework (SemFlow) and model them as a pair of reverse problems. Specifically, motivated by rectified flow theory, we train an ordinary differential equation (ODE) model to transport between the distributions of real images and semantic masks. As the training object is symmetric, samples belonging to the two distributions, images and semantic masks, can be effortlessly transferred reversibly. For semantic segmentation, our approach solves the contradiction between the randomness of diffusion outputs and the uniqueness of segmentation results. For image synthesis, we propose a finite perturbation approach to enhance the diversity of generated results without changing the semantic categories. Experiments show that our SemFlow achieves competitive results on semantic segmentation and semantic image synthesis tasks. We hope this simple framework will motivate people to rethink the unification of low-level and high-level vision.

Generic Object Tracking (GOT) is the problem of tracking target objects, specified by bounding boxes in the first frame of a video. While the task has received much attention in the last decades, researchers have almost exclusively focused on the single object setting. However multiobject GOT poses its own challenges and is more attractive in real-world applications. We attribute the lack of research interest into this problem to the absence of suitable benchmarks. In this work, we introduce a new largescale GOT benchmark, LaGOT, containing multiple annotated target objects per sequence. Our benchmark allows users to tackle key remaining challenges in GOT, aiming to increase robustness and reduce computation through joint tracking of multiple objects simultaneously. In addition, we propose a transformer-based GOT tracker baseline capable of joint processing of multiple objects through shared computation. Our approach achieves a 4× faster run-time in case of 10 concurrent objects compared to tracking each object independently and outperforms existing single object trackers on our new benchmark. In addition, our approach achieves highly competitive results on single-object GOT datasets, setting a new state of the art on TrackingNet with a success rate AUC of 84.4%. Our benchmark, code, results and trained models are available at https://github.com/visionml/pytracking.

Few-Shot Instance Segmentation (FSIS) requires detecting and segmenting novel classes with limited support examples. Existing methods based on Region Proposal Networks (RPNs) face two issues: 1) overfitting suppresses novel class objects and 2) dual-branch models require complex spatial correlation strategies to prevent spatial information loss when generating class prototypes. We introduce a unified framework, Reference Twice (RefT), to exploit the relationship between support and query features for FSIS and related tasks. Our three main contributions are: 1) a novel transformer-based baseline that avoids overfitting, offering a new direction for FSIS; 2) demonstrating that support object queries encode key factors after base training, allowing query features to be enhanced twice at both feature and query levels using simple cross-attention, thus avoiding complex spatial correlation interaction; and 3) introducing a class-enhanced base knowledge distillation loss to address the issue of DETR-like models struggling with incremental settings due to the input projection layer, enabling easy extension to incremental FSIS. Extensive experimental evaluations on the COCO dataset under three FSIS settings demonstrate that our method performs favorably against existing approaches across different shots, e.g., +8.2/+9.4 performance gain over state-of-the-art methods with 10/30-shots.

A desirable objective in self-supervised learning (SSL) is to avoid feature collapse. Whitening loss guarantees collapse avoidance by minimizing the distance between embeddings of positive pairs under the conditioning that the embeddings from different views are whitened. In this paper, we propose a framework with an informative indicator to analyze whitening loss, which provides a clue to demystify several interesting phenomena and a pivoting point connecting to other SSL methods. We show that batch whitening (BW) based methods do not impose whitening constraints on the embedding but only require the embedding to be full-rank. This full-rank constraint is also sufficient to avoid dimensional collapse. We further demonstrate that the stable rank of the embedding is invariant during training by gradient descent, given the assumption that embedding is updated with an infinitely small learning rate. Based on our analysis, we propose channel whitening with random group partition (CW-RGP), which exploits the advantages of BW-based methods in preventing collapse and avoids their disadvantages requiring large batch size. Experimental results on ImageNet classification and COCO object detection reveal that the proposed CW-RGP possesses a promising potential for learning good representations.

Although face swapping has attracted much attention in recent years, it remains a challenging problem. Existing methods leverage a large number of data samples to explore the intrinsic properties of face swapping without considering the semantic information of face images. Moreover, the representation of the identity information tends to be fixed, leading to suboptimal face swapping. In this paper, we present a simple yet efficient method named FaceSwapper, for one-shot face swapping based on Generative Adversarial Networks. Our method consists of a disentangled representation module and a semantic-guided fusion module. The disentangled representation module comprises an attribute encoder and an identity encoder, which aims to achieve the disentanglement of the identity and attribute information. The identity encoder is more flexible, and the attribute encoder contains more attribute details than its competitors. Benefiting from the disentangled representation, FaceSwapper can swap face images progressively. In addition, semantic information is introduced into the semantic-guided fusion module to control the swapped region and model the pose and expression more accurately. Experimental results show that our method achieves state-of-the-art results on benchmark datasets with fewer training samples.

Panoptic Part Segmentation (PPS) unifies panoptic and part segmentation into one task. Previous works utilize separate approaches to handle things, stuff, and part predictions without shared computation and task association. We aim to unify these tasks at the architectural level, designing the first end-to-end unified framework, Panoptic-PartFormer. Moreover, we find the previous metric PartPQ biases to PQ. To handle both issues, we first design a meta-architecture that decouples part features and things/stuff features, respectively. We model things, stuff, and parts as object queries and directly learn to optimize all three forms of prediction as a unified mask prediction and classification problem. We term our model as Panoptic-PartFormer. Second, we propose a new metric Part-Whole Quality (PWQ), better to measure this task from pixel-region and part-whole perspectives. It also decouples the errors for part segmentation and panoptic segmentation. Third, inspired by Mask2Former, based on our meta-architecture, we propose Panoptic-PartFormer++ and design a new part-whole cross-attention scheme to boost part segmentation qualities further. We design a new part-whole interaction method using masked cross attention. Finally, extensive ablation studies and analysis demonstrate the effectiveness of both Panoptic-PartFormer and Panoptic-PartFormer++. Compared with previous Panoptic-PartFormer, our Panoptic-PartFormer++ achieves 2% PartPQ and 3% PWQ improvements on the Cityscapes PPS dataset and 5% PartPQ on the Pascal Context PPS dataset. On both datasets, Panoptic-PartFormer++ achieves new state-of-the-art results. Our models can serve as a strong baseline and aid future research in PPS.

Video captioning aims to generate natural language descriptions for a given video clip. Existing methods mainly focus on end-to-end representation learning via word-by-word comparison between predicted captions and ground-truth texts. Although significant progress has been made, such supervised approaches neglect semantic alignment between visual and linguistic entities, which may negatively affect the generated captions. In this work, we propose a hierarchical modular network to bridge video representations and linguistic semantics at four granularities before generating captions: entity, verb, predicate, and sentence. Each level is implemented by one module to embed corresponding semantics into video representations. Additionally, we present a reinforcement learning module based on the scene graph of captions to better measure sentence similarity. Extensive experimental results show that the proposed method performs favorably against the state-of-the-art models on three widely-used benchmark datasets, including microsoft research video description corpus (MSVD), MSR-video to text (MSR-VTT), and video-and-TEXt (VATEX).

Optical aberration is a ubiquitous degeneration in realistic lens-based imaging systems. Optical aberrations are caused by the differences in the optical path length when light travels through different regions of the camera lens with different incident angles. The blur and chromatic aberrations manifest significant discrepancies when the optical system changes. This work designs a transferable and effective image simulation system of simple lenses via multi-wavelength, depth-aware, spatially-variant four-dimensional point spread functions (4D-PSFs) estimation by changing a small amount of lens-dependent parameters. The image simulation system can alleviate the overhead of dataset collecting and exploiting the principle of computational imaging for effective optical aberration correction. With the guidance of domain knowledge about the image formation model provided by the 4D-PSFs, we establish a multi-scale optical aberration correction network for degraded image reconstruction, which consists of a scene depth estimation branch and an image restoration branch. Specifically, we propose to predict adaptive filters with the depth-aware PSFs and carry out dynamic convolutions, which facilitate the model's generalization in various scenes. We also employ convolution and self-attention mechanisms for global and local feature extraction and realize a spatially-variant restoration. The multi-scale feature extraction complements the features across different scales and provides fine details and contextual features. Extensive experiments demonstrate that our proposed algorithm performs favorably against state-of-the-art restoration methods.

This article targets the task of novel category discovery (NCD), which aims to discover unknown categories when a certain number of classes are already known. The NCD task is challenging due to its closeness to real-world scenarios, where we have only encountered some partial classes and corresponding images. Unlike previous approaches to NCD, we propose a novel adaptive prototype learning method that leverages prototypes to emphasize category discrimination and alleviate the issue of missing annotations for novel classes. Concretely, the proposed method consists of two main stages: prototypical representation learning and prototypical self-training. In the first stage, we develop a robust feature extractor that could effectively handle images from both base and novel categories. This ability of instance and category discrimination of the feature extractor is boosted by self-supervised learning and adaptive prototypes. In the second stage, we utilize the prototypes again to rectify offline pseudo labels and train a final parametric classifier for category clustering. We conduct extensive experiments on four benchmark datasets, demonstrating our method's effectiveness and robustness with state-of-the-art performance.

Crowd counting is usually handled in a density map regression fashion, which is supervised via an L2 loss between the predicted density map and ground truth. To effectively regulate models, various improved L2 loss functions have been developed to find a better correspondence between predicted density and annotation positions. In this paper, we propose to predict the density map at one resolution but measure its quality via a derived log-formed loss at multiple resolutions. Unlike existing methods that assume density maps at different resolutions are independent, our loss is obtained by modeling the likelihood function inspired by the relationship of density maps across multi-resolutions. We find that the traditional single-resolution L2 loss is a particular case of our derived log-likelihood. We mathematically prove it is superior to a single-resolution L2 loss. Without bells and whistles, the proposed loss substantially improves several baselines and performs favorably compared to state-of-the-art methods on five crowd counting datasets: NWPU-Crowd, ShanghaiTech A & B, UCF-QNRF, and JHU-Crowd++. The source code and trained models are released athttps://github.com/streamer-AP/PML_Loss.git.

Blur artifacts can seriously degrade the visual quality of images, and numerous deblurring methods have been proposed for specific scenarios. However, in most real-world images, blur is caused by different factors, e.g., motion, and defocus. In this paper, we address how other deblurring methods perform in the case of multiple types of blur. For in-depth performance evaluation, we construct a new large-scale multi-cause image deblurring dataset (MC-Blur), including real-world and synthesized blurry images with different blur factors. The images in the proposed MC-Blur dataset are collected using other techniques: averaging sharp images captured by a 1000-fps high-speed camera, convolving Ultra-High-Definition (UHD) sharp images with large-size kernels, adding defocus to images, and real-world blurry images captured by various camera models. Based on the MC-Blur dataset, we conduct extensive benchmarking studies to compare SOTA methods in different scenarios, analyze their efficiency, and investigate the buildataset’s capacity. These benchmarking results provide a comprehensive overview of the advantages and limitations of current deblurring methods, revealing our dataset’s advances. The dataset is available to the public athttps://github.com/HDCVLab/MC-Blur-Dataset.

Owing to the success of transformer models, recent works study their applicability in 3D medical segmentation tasks. Within the transformer models, the self-attention mechanism is one of the main building blocks that strives to capture long-range dependencies, compared to the local convolutional-based design. However, the self-attention operation has quadratic complexity which proves to be a computational bottleneck, especially in volumetric medical imaging, where the inputs are 3D with numerous slices. In this paper, we propose a 3D medical image segmentation approach, named UNETR++, that offers both high-quality segmentation masks as well as efficiency in terms of parameters, compute cost, and inference speed. The core of our design is the introduction of a novel efficient paired attention (EPA) block that efficiently learns spatial and channel-wise discriminative features using a pair of inter-dependent branches based on spatial and channel attention. Our spatial attention formulation is efficient and has linear complexity with respect to the input. To enable communication between spatial and channel-focused branches, we share the weights of query and key mapping functions that provide a complimentary benefit (paired attention), while also reducing the complexity. Our extensive evaluations on five benchmarks, Synapse, BTCV, ACDC, BraTS, and Decathlon-Lung, reveal the effectiveness of our contributions in terms of both efficiency and accuracy. On Synapse, our UNETR++ sets a new state-of-the-art with a Dice Score of 87.2%, while significantly reducing parameters and FLOPs by over 71%, compared to the best method in the literature. Our code and models are available at: https://tinyurl.com/2p87x5xn.

While deep-learning-based tracking methods have achieved substantial progress, they entail large-scale and high-quality annotated data for sufficient training. To eliminate expensive and exhaustive annotation, we study self-supervised (SS) learning for visual tracking. In this work, we develop the crop-transform-paste operation, which is able to synthesize sufficient training data by simulating various appearance variations during tracking, including appearance variations of objects and background interference. Since the target state is known in all synthesized data, existing deep trackers can be trained in routine ways using the synthesized data without human annotation. The proposed target-aware data-synthesis method adapts existing tracking approaches within a SS learning framework without algorithmic changes. Thus, the proposed SS learning mechanism can be seamlessly integrated into existing tracking frameworks to perform training. Extensive experiments show that our method: 1) achieves favorable performance against supervised (Su) learning schemes under the cases with limited annotations; 2) helps deal with various tracking challenges such as object deformation, occlusion (OCC), or background clutter (BC) due to its manipulability; 3) performs favorably against the state-of-the-art unsupervised tracking methods; and 4) boosts the performance of various state-of-the-art Su learning frameworks, including SiamRPN++, DiMP, and TransT.

Given a piece of text, a video clip and a reference audio, the movie dubbing (also known as visual voice clone, V2C) task aims to generate speeches that match the speaker's emotion presented in the video using the desired speaker voice as reference. V2C is more challenging than conventional text-to-speech tasks as it additionally requires the generated speech to exactly match the varying emotions and speaking speed presented in the video. Unlike previous works, we propose a novel movie dubbing architecture to tackle these problems via hierarchical prosody modeling, which bridges the visual information to corresponding speech prosody from three aspects: lip, face, and scene. Specifically, we align lip movement to the speech duration, and convey facial expression to speech energy and pitch via attention mechanism based on valence and arousal representations inspired by the psychology findings. Moreover, we design an emotion booster to capture the atmosphere from global video scenes. All these embeddings are used together to generate mel-spectrogram, which is then converted into speech waves by an existing vocoder. Extensive experimental results on the V2C and Chem benchmark datasets demonstrate the favourable performance of the proposed method. The code and trained models will be made available at https://github.com/GalaxyCong/HPMDubbing

On a shutter press, modern handheld cameras capture multiple images in rapid succession and merge them to gen-erate a single image. However, individual frames in a burst are misaligned due to inevitable motions and contain mul-tiple degradations. The challenge is to properly align the successive image shots and merge their complimentary in-formation to achieve high-quality outputs. Towards this direction, we propose Burstormer: a novel transformer-based architecture for burst image restoration and enhancement. In comparison to existing works, our approach exploits multi-scale local and non-local features to achieve improved alignment and feature fusion. Our key idea is to enable inter-frame communication in the burst neighborhoods for information aggregation and progressive fusion while modeling the burst-wide context. However, the input burst frames need to be properly aligned before fusing their information. Therefore, we propose an enhanced de-formable alignment module for aligning burst features with regards to the reference frame. Unlike existing methods, the proposed alignment module not only aligns burst features but also exchanges fea-ture information and maintains focused communication with the reference frame through the proposed reference-based feature enrichment mechanism, which facilitates handling complex motions. After multi-level alignment and enrichment, we re-emphasize on inter-frame communication within burst using a cyclic burst sampling module. Finally, the inter-frame information is aggre-gated using the proposed burst feature fusion module followed by progressive upsampling. Our Burstormer outperforms state-of-the-art methods on burst super-resolution, burst denoising and burst low-light enhance-ment. Our codes and pre-trained models are available at https://github.com/akshaydudhane16/Burstormer.

Multi-modal- image-text models such as CLIP and LiT have demonstrated impressive performance on image classification benchmarks and their zero-shot generalization ability is particularly exciting. While the top-5 zero-shot accuracies of these models are very high, the top-1 accuracies are much lower (over 25% gap in some cases). We investigate the reasons for this performance gap and find that many of the failure cases are caused by ambiguity in the text prompts. First, we develop a simple and efficient zero-shot post-hoc method to identify images whose top-1 prediction is likely to be incorrect, by measuring consistency of the predictions w.r.t. multiple prompts and image transformations. We show that our procedure better predicts mistakes, outperforming the popular max logit baseline on selective prediction tasks. Next, we propose a simple and efficient way to improve accuracy on such uncertain images by making use of the WordNet hierarchy; specifically we augment the original class by incorporating its parent and children from the semantic label hierarchy, and plug the augmentation into text prompts. We conduct experiments on both CLIP and LiT models with five different ImageNet-based datasets. For CLIP, our method improves the top-1 accuracy by 17.13% on the uncertain subset and 3.6% on the entire ImageNet validation set. We also show that our method improves across ImageNet shifted datasets, four other datasets, and other model architectures such as LiT. The proposed method11Work carried out mainly at Google is hyperparameter-free, requires no additional model training and can be easily scaled to other large multi-modal architectures. Code is available at https://github.com/gyhandy/Hierarchy-CLIP.

Recently, AutoFlow has shown promising results on learning a training set for optical flow, but requires ground truth labels in the target domain to compute its search metric. Observing a strong correlation between the ground truth search metric and self-supervised losses, we introduce self-supervised AutoFlow to handle real-world videos without ground truth labels. Using self-supervised loss as the search metric, our self-supervised AutoFlow performs on par with AutoFlow on Sintel and KITTI where ground truth is available, and performs better on the real-world DAVIS dataset. We further explore using self-supervised AutoFlow in the (semi-)supervised setting and obtain competitive results against the state of the art.

Automatically estimating 3D skeleton, shape, camera viewpoints, and part articulation from sparse in-the-wild image ensembles is a severely under-constrained and challenging problem. Most prior methods rely on large-scale image datasets, dense temporal correspondence, or human annotations like camera pose, 2D keypoints, and shape templates. We propose Hi-LASSIE, which performs 3D articulated reconstruction from only 20–30 online images in the wild without any user-defined shape or skeleton templates. We follow the recent work of LASSIE that tackles a similar problem setting and make two significant advances. First, instead of relying on a manually annotated 3D skeleton, we automatically estimate a class-specific skeleton from the selected reference image. Second, we improve the shape reconstructions with novel instance-specific optimization strategies that allow reconstructions to faithful fit on each instance while preserving the class-specific priors learned across all images. Experiments on in-the-wild image ensembles show that Hi-LASSIE obtains higher fidelity state-of-the-art 3D reconstructions despite requiring minimum user input. Project page: chhankyao.github.io/hi-lassie/

Neural implicit surface representation methods show impressive reconstruction results but struggle to handle texture-less planar regions that widely exist in indoor scenes. Existing approaches addressing this leverage image prior that requires assistive networks trained with large-scale annotated datasets. In this work, we introduce a self-supervised super-plane constraint by exploring the free geometry cues from the predicted surface, which can further regularize the reconstruction of plane regions without any other ground truth annotations. Specifically, we introduce an iterative training scheme, where (i) grouping of pixels to formulate a super-plane (analogous to super-pixels), and (ii) optimizing of the scene reconstruction network via a super-plane constraint, are progressively conducted. We demonstrate that the model trained with superplanes surprisingly outperforms the one using conventional annotated planes, as individual super-plane statistically occupies a larger area and leads to more stable training. Extensive experiments show that our self-supervised super-plane constraint significantly improves 3D reconstruction quality even better than using ground truth plane segmentation. Additionally, the plane reconstruction results from our model can be used for auto-labeling for other vision tasks. The code and models are available at https://github.com/botaoye/S3PRecon.

We introduce the MAsked Generative VIdeo Transformer, MAGVIT, to tackle various video synthesis tasks with a single model. We introduce a 3D tokenizer to quantize a video into spatial-temporal visual tokens and propose an embedding method for masked video token modeling to facilitate multi-task learning. We conduct extensive experiments to demonstrate the quality, efficiency, and flexibility of MAGVIT. Our experiments show that (i) MAGVIT performs favorably against state-of-the-art approaches and establishes the best-published FVD on three video generation benchmarks, including the challenging Kinetics-600. (ii) MAGVIT outperforms existing methods in inference time by two orders of magnitude against diffusion models and by 60x against autoregressive models. (iii) A single MAGVIT model supports ten diverse generation tasks and generalizes across videos from different visual domains. The source code and trained models will be released to the public at https://magvit.cs.cmu.edu.

Weakly supervised video anomaly detection aims to identify abnormal events in videos using only video-level labels. Recently, two-stage self-training methods have achieved significant improvements by self-generating pseudo labels and self-refining anomaly scores with these labels. As the pseudo labels play a crucial role, we propose an enhancement framework by exploiting completeness and uncertainty properties for effective self-training. Specifically, we first design a multi-head classification module (each head serves as a classifier) with a diversity loss to maximize the distribution differences of predicted pseudo labels across heads. This encourages the generated pseudo labels to cover as many abnormal events as possible. We then devise an iterative uncertainty pseudo label refinement strategy, which improves not only the initial pseudo labels but also the updated ones obtained by the desired classifier in the second stage. Extensive experimental results demonstrate the proposed method performs favorably against state-of-the-art approaches on the UCF-Crime, TAD, and XD-Violence benchmark datasets.

This work focuses on training a single visual relationship detector predicting over the union of label spaces from multiple datasets. Merging labels spanning different datasets could be challenging due to inconsistent taxonomies. The issue is exacerbated in visual relationship detection when second-order visual semantics are introduced between pairs of objects. To address this challenge, we propose UniVRD, a novel bottom-up method for Unified Visual Relationship Detection by leveraging vision and language models (VLMs). VLMs provide well-aligned image and text embeddings, where similar relationships are optimized to be close to each other for semantic unification. Our bottom-up design enables the model to enjoy the benefit of training with both object detection and visual relationship datasets. Empirical results on both human-object interaction detection and scene-graph generation demonstrate the competitive performance of our model. UniVRD achieves 38.07 mAP on HICO-DET, outperforming the current best bottom-up HOI detector by 14.26 mAP. More importantly, we show that our unified detector performs as well as dataset-specific models in mAP, and achieves further improvements when we scale up the model. Our code will be made publicly available on GitHub1.

Online Action Detection (OAD) is the task of identifying actions in streaming videos without access to future frames. Much effort has been devoted to effectively capturing long-range dependencies, with transformers receiving the spotlight for their ability to capture long-range temporal structures. In contrast, RNNs have received less attention lately, due to their lower performance compared to recent methods that utilize transformers. In this paper, we investigate the underlying reasons for the inferior performance of RNNs compared to transformer-based algorithms. Our findings indicate that the discrepancy between training and inference is the primary hindrance to the effective training of RNNs. To address this, we propose applying non-uniform weights to the loss computed at each time step, which allows the RNN model to learn from the predictions made in an environment that better resembles the inference stage. Extensive experiments on three benchmark datasets, THUMOS, TVSeries, and FineAction demonstrate that a minimal RNN-based model trained with the proposed methodology performs equally or better than the existing best methods with a significant increase in efficiency. The code is available at https://github.com/jbistanbul/MiniROAD.

Continual learning aims to emulate the human ability to continually accumulate knowledge over sequential tasks. The main challenge is to maintain performance on previously learned tasks after learning new tasks, i.e., to avoid catastrophic forgetting. We propose a Channel-wise Lightweight Reprogramming (CLR) approach that helps convolutional neural networks (CNNs) overcome catastrophic forgetting during continual learning. We show that a CNN model trained on an old task (or self-supervised proxy task) could be "reprogrammed" to solve a new task by using our proposed lightweight (very cheap) reprogramming parameter. With the help of CLR, we have a better stability-plasticity trade-off to solve continual learning problems: To maintain stability and retain previous task ability, we use a common task-agnostic immutable part as the shared "anchor" parameter set. We then add task-specific lightweight reprogramming parameters to reinterpret the outputs of the immutable parts, to enable plasticity and integrate new knowledge. To learn sequential tasks, we only train the lightweight reprogramming parameters to learn each new task. Reprogramming parameters are task-specific and exclusive to each task, which makes our method immune to catastrophic forgetting. To minimize the parameter requirement of reprogramming to learn new tasks, we make reprogramming lightweight by only adjusting essential kernels and learning channel-wise linear mappings from anchor parameters to task-specific domain knowledge. We show that, for general CNNs, the CLR parameter increase is less than 0.6% for any new task. Our method outperforms 13 state-of-the-art continual learning baselines on a new challenging sequence of 53 image classification datasets. Code and data are in the top link.

Recent advances of monocular 3D object detection facilitate the 3D multi-object tracking task based on lowcost camera sensors. In this paper, we find that the motion cue of objects along different time frames is critical in 3D multi-object tracking, which is less explored in existing monocular-based approaches. To this end, we propose MoMA-M3T, a framework that mainly consists of three motion-aware components. First, we represent the possible movement of an object related to all object tracklets in the feature space as its motion features. Then, we further model the historical object tracklet along the time frame in a spatial-temporal perspective via a motion transformer. Finally, we propose a motion-aware matching module to associate historical object tracklets and current observations as final tracking results. We conduct extensive experiments on the nuScenes and KITTI datasets to demonstrate that our MoMA-M3T achieves competitive performance against state-of-the-art methods. Moreover, the proposed tracker is flexible and can be easily plugged into existing image-based 3D object detectors without re-training. Code and models are available at https://github.com/kuanchihhuang/MoMA-M3T.

Crowd counting has recently attracted significant attention in the field of computer vision due to its wide applications to image understanding. Numerous methods have been proposed and achieved state-of-the-art performance for real-world tasks. However, existing approaches do not perform well under adverse weather such as haze, rain, and snow since the visual appearances of crowds in such scenes are drastically different from those images in clear weather of typical datasets. In this paper, we propose a method for robust crowd counting in adverse weather scenarios. Instead of using a two-stage approach that involves image restoration and crowd counting modules, our model learns effective features and adaptive queries to account for large appearance variations. With these weather queries, the proposed model can learn the weather information according to the degradation of the input image and optimize with the crowd counting module simultaneously. Experimental results show that the proposed algorithm is effective in counting crowds under different weather types on benchmark datasets. The source code is available in our project page.

Prompt learning has emerged as an efficient alternative for fine-tuning foundational models, such as CLIP, for various downstream tasks. Conventionally trained using the task-specific objective, i.e., cross-entropy loss, prompts tend to overfit downstream data distributions and find it challenging to capture task-agnostic general features from the frozen CLIP. This leads to the loss of the model’s original generalization capability. To address this issue, our work introduces a self-regularization framework for prompting called PromptSRC (Prompting with Self-regulating Constraints). PromptSRC guides the prompts to optimize for both task-specific and task-agnostic general representations using a three-pronged approach by: (a) regulating prompted representations via mutual agreement maximization with the frozen model, (b) regulating with self-ensemble of prompts over the training trajectory to encode their complementary strengths, and (c) regulating with textual diversity to mitigate sample diversity imbalance with the visual branch. To the best of our knowledge, this is the first regularization framework for prompt learning that avoids overfitting by jointly attending to pre-trained model features, the training trajectory during prompting, and the textual diversity. PromptSRC explicitly steers the prompts to learn a representation space that maximizes performance on downstream tasks without compromising CLIP generalization. We perform extensive experiments on 4 benchmarks where PromptSRC overall performs favorably well compared to the existing methods. Our code and pre-trained models are publicly available at: https://github.com/muzairkhattak/PromptSRC.

We present a method to efficiently generate 3D-aware high-resolution images that are view-consistent across multiple target views. The proposed multiplane neural radiance model, named GMNR, consists of a novel α-guided view-dependent representation (α-VdR) module for learning view-dependent information. The α-VdR module, faciliated by an α-guided pixel sampling technique, computes the view-dependent representation efficiently by learning viewing direction and position coefficients. Moreover, we propose a view-consistency loss to enforce photometric similarity across multiple views. The GMNR model can generate 3D-aware high-resolution images that are view-consistent across multiple camera poses, while maintaining the computational efficiency in terms of both training and inference time. Experiments on three datasets demonstrate the effectiveness of the proposed modules, leading to favorable results in terms of both generation quality and inference time, compared to existing approaches. Our GMNR model generates 3D-aware images of 1024×1024 pixels with 17.6 FPS on a single V100. Code: https://github.com/VIROBO-15/GMNR

Existing visual tracking methods typically take an image patch as the reference of the target to perform tracking. However, a single image patch cannot provide a complete and precise concept of the target object as images are limited in their ability to abstract and can be ambiguous, which makes it difficult to track targets with drastic variations. In this paper, we propose the CiteTracker to enhance target modeling and inference in visual tracking by connecting images and text. Specifically, we develop a text generation module to convert the target image patch into a descriptive text containing its class and attribute information, providing a comprehensive reference point for the target. In addition, a dynamic description module is designed to adapt to target variations for more effective target representation. We then associate the target description and the search image using an attention-based correlation module to generate the correlated features for target state reference. Extensive experiments on five diverse datasets are conducted to evaluate the proposed algorithm and the favorable performance against the state-of-the-art methods demonstrates the effectiveness of the proposed tracking method. The source code and trained models will be released at https://github.com/NorahGreen/CiteTracker.

Toward infinite-scale 3D city synthesis, we propose a novel framework, InfiniCity, which constructs and renders an unconstrainedly large and 3D-grounded environment from random noises. InfiniCity decomposes the seemingly impractical task into three feasible modules, taking advantage of both 2D and 3D data. First, an infinite-pixel image synthesis module generates arbitrary-scale 2D maps from the bird’s-eye view. Next, an octree-based voxel completion module lifts the generated 2D map to 3D octrees. Finally, a voxel-based neural rendering module texturizes the voxels and renders 2D images. InfiniCity can thus synthesize arbitrary-scale and traversable 3D city environments. We quantitatively and qualitatively demonstrate the efficacy of the proposed framework.

Dense image segmentation tasks (e.g., semantic, panoptic) are useful for image editing, but existing methods can hardly generalize well in an in-the-wild setting where there are unrestricted image domains, classes, and image resolution & quality variations. Motivated by these observations, we construct a new entity segmentation dataset, with a strong focus on high-quality dense segmentation in the wild. The dataset contains images spanning diverse image domains and entities, along with plentiful high-resolution images and high-quality mask annotations for training and testing. Given the high-quality and -resolution nature of the dataset, we propose CropFormer which is designed to tackle the intractability of instance-level segmentation on high-resolution images. It improves mask prediction by fusing high-res image crops that provides more fine-grained image details and the full image. CropFormer is the first query-based Transformer architecture that can effectively fuse mask predictions from multiple image views, by learning queries that effectively associate the same entities across the full image and its crop. With CropFormer, we achieve a significant AP gain of 1.9 on the challenging entity segmentation task. Furthermore, CropFormer consistently improves the accuracy of traditional segmentation tasks and datasets. The dataset and code are released at http://luqi.info/entityv2.github.io/.

Self-attention has become a defacto choice for capturing global context in various vision applications. However, its quadratic computational complexity with respect to image resolution limits its use in real-time applications, especially for deployment on resource-constrained mobile devices. Although hybrid approaches have been proposed to combine the advantages of convolutions and self-attention for a better speed-accuracy trade-off, the expensive matrix multiplication operations in self-attention remain a bottleneck. In this work, we introduce a novel efficient additive attention mechanism that effectively replaces the quadratic matrix multiplication operations with linear element-wise multiplications. Our design shows that the key-value interaction can be replaced with a linear layer without sacrificing any accuracy. Unlike previous state-of-the-art methods, our efficient formulation of self-attention enables its usage at all stages of the network. Using our proposed efficient additive attention, we build a series of models called "SwiftFormer" which achieves state-of-the-art performance in terms of both accuracy and mobile inference speed. Our small variant achieves 78.5% top-1 ImageNet-1K accuracy with only 0.8 ms latency on iPhone 14, which is more accurate and 2× faster compared to MobileViT-v2. Our code and models: https://tinyurl.com/5ft8v46w

Recent novel view synthesis methods obtain promising results for relatively small scenes, e.g., indoor environments and scenes with a few objects, but tend to fail for unbounded outdoor scenes with a single image as input. In this paper, we introduce SAMPLING, a Scene-adaptive Hierarchical Multiplane Images Representation for Novel View Synthesis from a Single Image based on improved multiplane images (MPI). Observing that depth distribution varies significantly for unbounded outdoor scenes, we employ an adaptive-bins strategy for MPI to arrange planes in accordance with each scene image. To represent intricate geometry and multi-scale details, we further introduce a hierarchical refinement branch, which results in high-quality synthesized novel views. Our method demonstrates considerable performance gains in synthesizing large-scale unbounded outdoor scenes using a single image on the KITTI dataset and generalizes well to the unseen Tanks and Temples dataset. The code and models will be made available at https://pkuvdig.github.io/SAMPLING/.

We present Muse, a text-to-image Transformermodel that achieves state-of-the-art image genera-tion performance while being significantly moreefficient than diffusion or autoregressive models.Muse is trained on a masked modeling task indiscrete token space: given the text embeddingextracted from a pre-trained large language model(LLM), Muse learns to predict randomly maskedimage tokens. Compared to pixel-space diffusionmodels, such as Imagen and DALL-E 2, Muse issignificantly more efficient due to the use of dis-crete tokens and requires fewer sampling itera-tions; compared to autoregressive models such asParti, Muse is more efficient due to the use of par-allel decoding. The use of a pre-trained LLM en-ables fine-grained language understanding, whichtranslates to high-fidelity image generation andthe understanding of visual concepts such as ob-jects, their spatial relationships, pose, cardinalityetc. Our 900M parameter model achieves a newSOTA on CC3M, with an FID score of 6.06. TheMuse 3B parameter model achieves an FID of7.88 on zero-shot COCO evaluation, along with aCLIP score of 0.32. Muse also directly enables anumber of image editing applications without theneed to fine-tune or invert the model: inpainting,outpainting, and mask-free editing. More resultsand videos demonstrating editing are available at https://muse-icml.github.io/

Recent studies have shown that paddings in convolutional neural networks encode absolute position information which can negatively affect the model performance for certain tasks. However, existing metrics for quantifying the strength of positional information remain unreliable and frequently lead to erroneous results. To address this issue, we propose novel metrics for measuring and visualizing the encoded positional information. We formally define the encoded information as Position-information Pattern from Padding (PPP) and conduct a series of experiments to study its properties as well as its formation. The proposed metrics measure the presence of positional information more reliably than the existing metrics based on PosENet and tests in F-Conv. We also demonstrate that for any extant (and proposed) padding schemes, PPP is primarily a learning artifact and is less dependent on the characteristics of the underlying padding schemes.

Pre-trained multimodal foundation models have demonstrated remarkable generalizability but pose challenges for deployment due to their large sizes. One effective approach to reducing their sizes is layerwise distillation, wherein small student models are trained to match the hidden representations of large teacher models at each layer. Motivated by our observation that certain architecture components, referred to as modules, contribute more significantly to the student's performance than others, we propose to track the contributions of individual modules by recording the loss decrement after distillation each module and choose the module with a greater contribution to distill more frequently. Such an approach can be naturally formulated as a multi-armed bandit (MAB) problem, where modules and loss decrements are considered as arms and rewards, respectively. We then develop a modified-Thompson sampling algorithm named OPTIMA to address the nonstationarity of module contributions resulting from model updating. Specifically, we leverage the observed contributions in recent history to estimate the changing contribution of each module and select modules based on these estimations to maximize the cumulative contribution. We evaluate the effectiveness of OPTIMA through distillation experiments on various multimodal understanding and image captioning tasks, using the CoCa-Large model \citep{yu2022coca} as the teacher model.

Semi-supervised object detection is crucial for 3D scene understanding, efficiently addressing the limitation of acquiring large-scale 3D bounding box annotations. Existing methods typically employ a teacher-student framework with pseudo-labeling to leverage unlabeled point clouds. However, producing reliable pseudo-labels in a diverse 3D space still remains challenging. In this work, we propose Diffusion-SS3D, a new perspective of enhancing the quality of pseudo-labels via the diffusion model for semi-supervised 3D object detection. Specifically, we include noises to produce corrupted 3D object size and class label distributions, and then utilize the diffusion model as a denoising process to obtain bounding box outputs. Moreover, we integrate the diffusion model into the teacher-student framework, so that the denoised bounding boxes can be used to improve pseudo-label generation, as well as the entire semi-supervised learning process. We conduct experiments on the ScanNet and SUN RGB-D benchmark datasets to demonstrate that our approach achieves state-of-the-art performance against existing methods. We also present extensive analysis to understand how our diffusion model design affects performance in semi-supervised learning. The source code will be available at https://github.com/luluho1208/Diffusion-SS3D.

In this paper, we present a novel problem, namely video timeline modeling. Our objective is to create a video-associated timeline from a set of videos related to a specific topic, thereby facilitating the content and structure understanding of the story being told. This problem has significant potential in various real-world applications, for instance, news story summarization. To bootstrap research in this area, we curate a realistic benchmark dataset, YouTube-News-Timeline, consisting of over $12$k timelines and $300$k YouTube news videos. Additionally, we propose a set of quantitative metrics to comprehensively evaluate and compare methodologies. With such a testbed, we further develop and benchmark several deep learning approaches to tackling this problem. We anticipate that this exploratory work will pave the way for further research in video timeline modeling. The assets are available via https://github.com/google-research/google-research/tree/master/video_timeline_modeling.

Despite the progress of image segmentation for accurate visual entity segmentation, completing the diverse requirements of image editing applications for different-level region-of-interest selections remains unsolved. In this paper, we propose a new task, All-Inclusive Multi-Level Segmentation (AIMS), which segments visual regions into three levels: part, entity, and relation (two entities with some semantic relationships). We also build a unified AIMS model through multi-dataset multi-task training to address the two major challenges of annotation inconsistency and task correlation. Specifically, we propose task complementarity, association, and prompt mask encoder for three-level predictions. Extensive experiments demonstrate the effectiveness and generalization capacity of our method compared to other state-of-the-art methods on a single dataset or the concurrent work on segment anything. We will make our code and training model publicly available.

Estimating 3D articulated shapes like animal bodies from monocular images is inherently challenging due to the ambiguities of camera viewpoint, pose, texture, lighting, etc. We propose ARTIC3D, a self-supervised framework to reconstruct per-instance 3D shapes from a sparse image collection in-the-wild. Specifically, ARTIC3D is built upon a skeleton-based surface representation and is further guided by 2D diffusion priors from Stable Diffusion. First, we enhance the input images with occlusions/truncation via 2D diffusion to obtain cleaner mask estimates and semantic features. Second, we perform diffusion-guided 3D optimization to estimate shape and texture that are of high-fidelity and faithful to input images. We also propose a novel technique to calculate more stable image-level gradients via diffusion models compared to existing alternatives. Finally, we produce realistic animations by fine-tuning the rendered shape and texture under rigid part transformations. Extensive evaluations on multiple existing datasets as well as newly introduced noisy web image collections with occlusions and truncation demonstrate that ARTIC3D outputs are more robust to noisy images, higher quality in terms of shape and texture details, and more realistic when animated.

In this work, we introduce Semantic Pyramid AutoEncoder (SPAE) for enabling frozen LLMs to perform both understanding and generation tasks involving non-linguistic modalities such as images or videos. SPAE converts between raw pixels and interpretable lexical tokens (or words) extracted from the LLM's vocabulary. The resulting tokens capture both the rich semantic meaning and the fine-grained details needed for visual reconstruction, effectively translating the visual content into a language comprehensible to the LLM, and empowering it to perform a wide array of multimodal tasks. Our approach is validated through in-context learning experiments with frozen PaLM 2 and GPT 3.5 on a diverse set of image understanding and generation tasks. Our method marks the first successful attempt to enable a frozen LLM to generate image content while surpassing state-of-the-art performance in image understanding tasks, under the same setting, by over 25%.

Text-to-image diffusion models have made significant advances in generating and editing high-quality images. As a result, numerous approaches have explored the ability of diffusion model features to understand and process single images for downstream tasks, e.g., classification, semantic segmentation, and stylization. However, significantly less is known about what these features reveal across multiple, different images and objects. In this work, we exploit Stable Diffusion (SD) features for semantic and dense correspondence and discover that with simple post-processing, SD features can perform quantitatively similar to SOTA representations. Interestingly, the qualitative analysis reveals that SD features have very different properties compared to existing representation learning features, such as the recently released DINOv2: while DINOv2 provides sparse but accurate matches, SD features provide high-quality spatial information but sometimes inaccurate semantic matches. We demonstrate that a simple fusion of these two features works surprisingly well, and a zero-shot evaluation using nearest neighbors on these fused features provides a significant performance gain over state-of-the-art methods on benchmark datasets, e.g., SPair-71k, PF-Pascal, and TSS. We also show that these correspondences can enable interesting applications such as instance swapping in two images. Project page: https://sd-complements-dino.github.io/.

Generative Adversarial Networks (GANs) typically suffer from overfitting when limited training data is available. To facilitate GAN training, current methods propose to use data-specific augmentation techniques. Despite the effectiveness, it is difficult for these methods to scale to practical applications. In this article, we present ScoreMix, a novel and scalable data augmentation approach for various image synthesis tasks. We first produce augmented samples using the convex combinations of the real samples. Then, we optimize the augmented samples by minimizing the norms of the data scores, i.e., the gradients of the log-density functions. This procedure enforces the augmented samples close to the data manifold. To estimate the scores, we train a deep estimation network with multi-scale score matching. For different image synthesis tasks, we train the score estimation network using different data. We do not require the tuning of the hyperparameters or modifications to the network architecture. The ScoreMix method effectively increases the diversity of data and reduces the overfitting problem. Moreover, it can be easily incorporated into existing GAN models with minor modifications. Experimental results on numerous tasks demonstrate that GAN models equipped with the ScoreMix method achieve significant improvements.

Empowered by large datasets, e.g., ImageNet and MS COCO, unsupervised learning on large-scale data has enabled significant advances for classification tasks. However, whether the large-scale unsupervised semantic segmentation can be achieved remains unknown. There are two major challenges: i) we need a large-scale benchmark for assessing algorithms; ii) we need to develop methods to simultaneously learn category and shape representation in an unsupervised manner. In this work, we propose a new problem of large-scale unsupervised semantic segmentation (LUSS) with a newly created benchmark dataset to help the research progress. Building on the ImageNet dataset, we propose the ImageNet-S dataset with 1.2 million training images and 50k high-quality semantic segmentation annotations for evaluation. Our benchmark has a high data diversity and a clear task objective. We also present a simple yet effective method that works surprisingly well for LUSS. In addition, we benchmark related un/weakly/fully supervised methods accordingly, identifying the challenges and possible directions of LUSS. The benchmark and source code is publicly available at https://github.com/LUSSeg.

Transformer models have achieved outstanding results on a variety of language tasks, such as text classification, ma- chine translation, and question answering. This success in the field of Natural Language Processing (NLP) has sparked interest in the computer vision community to apply these models to vision and multi-modal learning tasks. However, visual data has a unique structure, requiring the need to rethink network designs and training methods. As a result, Transformer models and their variations have been suc- cessfully used for image recognition, object detection, seg- mentation, image super-resolution, video understanding, image generation, text-image synthesis, and visual question answering, among other applications.

Facial Attribute Manipulation (FAM) aims to aesthetically modify a given face image to render desired attributes, which has received significant attention due to its broad practical applications ranging from digital entertainment to biometric forensics. In the last decade, with the remarkable success of Generative Adversarial Networks (GANs) in synthesizing realistic images, numerous GAN-based models have been proposed to solve FAM with various problem formulation approaches and guiding information representations. This paper presents a comprehensive survey of GAN-based FAM methods with a focus on summarizing their principal motivations and technical details. The main contents of this survey include: (i) an introduction to the research background and basic concepts related to FAM, (ii) a systematic review of GAN-based FAM methods in three main categories, and (iii) an in-depth discussion of important properties of FAM methods, open issues, and future research directions. This survey not only builds a good starting point for researchers new to this field but also serves as a reference for the vision community.

We present compact and effective deep convolutional neural networks (CNNs) by exploring properties of videos for video deblurring. Motivated by the non-uniform blur property that not all the pixels of the frames are blurry, we develop a CNN to integrate a temporal sharpness prior (TSP) for removing blur in videos. The TSP exploits sharp pixels from adjacent frames to facilitate the CNN for better frame restoration. Observing that the motion field is related to latent frames instead of blurry ones in the image formation model, we develop an effective cascaded training approach to solve the proposed CNN in an end-to-end manner. As videos usually contain similar contents within and across frames, we propose a non-local similarity mining approach based on a self-attention method with the propagation of global features to constrain CNNs for frame restoration. We show that exploring the domain knowledge of videos can make CNNs more compact and efficient, where the CNN with the non-local spatial-temporal similarity is 3× smaller than the state-of-the-art methods in terms of model parameters while its performance gains are at least 1 dB higher in terms of PSNRs. Extensive experimental results show that our method performs favorably against state-of-the-art approaches on benchmarks and real-world videos.

The softmax cross-entropy loss function has been widely used to train deep models for various tasks. In this work, we propose a Gaussian mixture (GM) loss function for deep neural networks for visual classification. Unlike the softmax cross-entropy loss, our method explicitly shapes the deep feature space towards a Gaussian Mixture distribution. With a classification margin and a likelihood regularization, the GM loss facilitates both high classification performance and accurate modeling of the feature distribution. The GM loss can be readily used to distinguish the adversarial examples based on the discrepancy between feature distributions of clean and adversarial examples. Furthermore, theoretical analysis shows that a symmetric feature space can be achieved by using the GM loss, which enables the models to perform robustly against adversarial attacks. The proposed model can be implemented easily and efficiently without introducing more trainable parameters. Extensive evaluations demonstrate that the method with the GM loss performs favorably on image classification, face recognition, and detection as well as recognition of adversarial examples generated by various attacks.

GAN inversion aims to invert a given image back into the latent space of a pretrained GAN model so that the image can be faithfully reconstructed from the inverted code by the generator. As an emerging technique to bridge the real and fake image domains, GAN inversion plays an essential role in enabling pretrained GAN models, such as StyleGAN and BigGAN, for applications of real image editing. Moreover, GAN inversion interprets GAN's latent space and examines how realistic images can be generated. In this paper, we provide a survey of GAN inversion with a focus on its representative algorithms and its applications in image restoration and image manipulation. We further discuss the trends and challenges for future research. A curated list of GAN inversion methods, datasets, and other related information can be found at https://github.com/weihaox/awesome-gan-inversion.

Given a degraded input image, image restoration aims to recover the missing high-quality image content. Numerous applications demand effective image restoration, e.g., computational photography, surveillance, autonomous vehicles, and remote sensing. Significant advances in image restoration have been made in recent years, dominated by convolutional neural networks (CNNs). The widely-used CNN-based methods typically operate either on full-resolution or on progressively low-resolution representations. In the former case, spatial details are preserved but the contextual information cannot be precisely encoded. In the latter case, generated outputs are semantically reliable but spatially less accurate. This paper presents a new architecture with a holistic goal of maintaining spatially-precise high-resolution representations through the entire network, and receiving complementary contextual information from the low-resolution representations. The core of our approach is a multi-scale residual block containing the following key elements: (a) parallel multi-resolution convolution streams for extracting multi-scale features, (b) information exchange across the multi-resolution streams, (c) non-local attention mechanism for capturing contextual information, and (d) attention based multi-scale feature aggregation. Our approach learns an enriched set of features that combines contextual information from multiple scales, while simultaneously preserving the high-resolution spatial details. Extensive experiments on six real image benchmark datasets demonstrate that our method, named as MIRNet-v2, achieves state-of-the-art results for a variety of image processing tasks, including defocus deblurring, image denoising, super-resolution, and image enhancement. The source code and pre-trained models are available at https://github.com/swz30/MIRNetv2.

The study of mouse social behaviors has been increasingly undertaken in neuroscience research. However, automated quantification of mouse behaviors from the videos of interacting mice is still a challenging problem, where object tracking plays a key role in locating mice in their living spaces. Artificial markers are often applied for multiple mice tracking, which are intrusive and consequently interfere with the movements of mice in a dynamic environment. In this article, we propose a novel method to continuously track several mice and individual parts without requiring any specific tagging. First, we propose an efficient and robust deep-learning-based mouse part detection scheme to generate part candidates. Subsequently, we propose a novel Bayesian-inference integer linear programming (BILP) model that jointly assigns the part candidates to individual targets with necessary geometric constraints while establishing pair-wise association between the detected parts. There is no publicly available dataset in the research community that provides a quantitative test bed for part detection and tracking of multiple mice, and we here introduce a new challenging Multi-Mice PartsTrack dataset that is made of complex behaviors. Finally, we evaluate our proposed approach against several baselines on our new datasets, where the results show that our method outperforms the other state-of-the-art approaches in terms of accuracy. We also demonstrate the generalization ability of the proposed approach on tracking zebra and locust.

Image outpainting seeks for a semantically consistent extension of the input image beyond its available content. Compared to inpainting - filling in missing pixels in a way coherent with the neighboring pixels - outpainting can be achieved in more diverse ways since the problem is less constrained by the surrounding pixels. Existing image outpainting methods pose the problem as a conditional image-to-image translation task, often generating repetitive structures and textures by replicating the content available in the input image. In this work, we formulate the problem from the perspective of inverting generative adversarial networks. Our generator renders micro-patches conditioned on their joint latent code as well as their individual positions in the image. To outpaint an image, we seek for multiple latent codes not only recovering available patches but also synthesizing diverse outpainting by patch-based generation. This leads to richer structure and content in the outpainted regions. Furthermore, our formulation allows for outpainting conditioned on the categorical input, thereby enabling flexible user controls. Extensive experimental results demonstrate the proposed method performs favorably against existing in- and outpainting methods, featuring higher visual quality and diversity.

Modern handheld devices can acquire burst image sequence in a quick succession. However, the individual acquired frames suffer from multiple degradations and are misaligned due to camera shake and object motions. The goal of Burst Image Restoration is to effectively combine complimentary cues across multiple burst frames to generate high-quality outputs. Towards this goal, we develop a novel approach by solely focusing on the effective information exchange between burst frames, such that the degradations get filtered out while the actual scene details are preserved and enhanced. Our central idea is to create a set of pseudo-burst features that combine complimentary information from all the input burst frames to seamlessly exchange information. However, the pseudo-burst cannot be successfully created unless the individual burst frames are properly aligned to discount inter-frame movements. Therefore, our approach initially extracts pre-processed features from each burst frame and matches them using an edge-boosting burst alignment module. The pseudo-burst features are then created and enriched using multi-scale contextual information. Our final step is to adaptively aggregate information from the pseudo-burst features to progressively increase resolution in multiple stages while merging the pseudo-burst features. In comparison to existing works that usually follow a late fusion scheme with single-stage upsampling, our approach performs favorably, delivering state-of-the-art performance on burst super-resolution, burst low-light image enhancement and burst denoising tasks. The source code and pre-trained models are available at https://github.com/akshaydudhane16/BIPNet.

Existing methods for video interpolation heavily rely on deep convolution neural networks, and thus suffer from their intrinsic limitations, such as content-agnostic kernel weights and restricted receptive field. To address these issues, we propose a Transformer-based video interpolation framework that allows content-aware aggregation weights and considers long-range dependencies with the self-attention operations. To avoid the high computational cost of global self-attention, we introduce the concept of local attention into video interpolation and extend it to the spatial-temporal domain. Furthermore, we propose a space-time separation strategy to save memory usage, which also improves performance. In addition, we develop a multi-scale frame synthesis scheme to fully realize the potential of Transformers. Extensive experiments demonstrate the proposed model performs favorably against the state-of-the-art methods both quantitatively and qualitatively on a variety of benchmark datasets. The code and models are released at https://github.com/zhshi0816/Video-Frame-Interpolation-Transformer.

Video captioning aims to generate natural language descriptions according to the content, where representation learning plays a crucial role. Existing methods are mainly developed within the supervised learning framework via word-by-word comparison of the generated caption against the ground-truth text without fully exploiting linguistic semantics. In this work, we propose a hierarchical modular network to bridge video representations and linguistic semantics from three levels before generating captions. In particular, the hierarchy is composed of: (I) Entity level, which highlights objects that are most likely to be mentioned in captions. (II) Predicate level, which learns the actions conditioned on highlighted objects and is supervised by the predicate in captions. (III) Sentence level, which learns the global semantic representation and is supervised by the whole caption. Each level is implemented by one module. Extensive experimental results show that the proposed method performs favorably against the state-of-the-art models on the two widely-used benchmarks: MSVD 104.0% and MSR-VTT 51.5% in CIDEr score. Code will be made available at https://github.com/MarcusNerva/HMN.

Modern self-supervised learning algorithms typically enforce persistency of instance representations across views. While being very effective on learning holistic image and video representations, such an objective becomes suboptimal for learning spatio-temporally fine-grained features in videos, where scenes and instances evolve through space and time. In this paper, we present Contextualized Spatio-Temporal Contrastive Learning (ConST-CL) to effectively learn spatio-temporally fine-grained video representations via self-supervision. We first design a region-based pretext task which requires the model to transform instance representations from one view to another, guided by context features. Further, we introduce a simple network design that successfully reconciles the simultaneous learning process of both holistic and local representations. We evaluate our learned representations on a variety of downstream tasks and show that ConST-CL achieves competitive results on 6 datasets, including Kinetics, UCF, HMDB, AVA-Kinetics, AVA and OTB. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/projects/const_cl.

Since convolutional neural networks (CNNs) perform well at learning generalizable image priors from large-scale data, these models have been extensively applied to image restoration and related tasks. Recently, another class of neural architectures, Transformers, have shown significant performance gains on natural language and high-level vision tasks. While the Transformer model mitigates the shortcomings of CNNs (i.e., limited receptive field and inadaptability to input content), its computational complexity grows quadratically with the spatial resolution, therefore making it infeasible to apply to most image restoration tasks involving high-resolution images. In this work, we propose an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images. Our model, named Restoration Transformer (Restormer), achieves state-of-the-art results on several image restoration tasks, including image deraining, single-image motion deblurring, defocus deblurring (single-image and dual-pixel data), and image denoising (Gaussian grayscale/color denoising, and real image denoising). The source code and pre-trained models are available at https://github.com/swz30/Restormer.

Predicting multiple future frames from a given video is a challenging problem due to several factors such as changing camera, dynamically moving objects, occlusions, etc. While recent deep learning methods have made significant progress on the video prediction problem, most methods predict the immediate or a fixed number of future frames. To obtain longer-term frame predictions, existing techniques usually process the predicted frames iteratively, resulting in blurry or inconsistent predictions. In this work, we present a new approach that can predict an arbitrary number of future video frames with a single forward pass through the network. Instead of directly predicting a fixed number of future optical flows or frames, we learn temporal motion encodings, i.e., temporal motion basis vectors and a network to predict the coefficients. The learned motion basis can be easily extended to arbitrary length at inference time, enabling us to predict an arbitrary number of future frames. Experiments on benchmark datasets show that our approach performs favorably against several competitive techniques even for the next frame prediction setting. When evaluated under 5-frame or 10-frame prediction settings, the proposed method achieves higher performance gains over the state-of-the-art techniques that iteratively process the predictions.

Creating high-quality articulated 3D models of animals is challenging either via manual creation or using 3D scanning tools. Therefore, techniques to reconstruct articulated 3D objects from 2D images are crucial and highly useful. In this work, we propose a practical problem setting to estimate 3D pose and shape of animals given only a few (10-30) in-the-wild images of a particular animal species (say, horse). Contrary to existing works that rely on pre-defined template shapes, we do not assume any form of 2D or 3D ground-truth annotations, nor do we leverage any multi-view or temporal information. Moreover, each input image ensemble can contain animal instances with varying poses, backgrounds, illuminations, and textures. Our key insight is that 3D parts have much simpler shape compared to the overall animal and that they are robust w.r.t. animal pose articulations. Following these insights, we propose LASSIE, a novel optimization framework which discovers 3D parts in a self-supervised manner with minimal user intervention. A key driving force behind LASSIE is the enforcing of 2D-3D part consistency using self-supervisory deep features. Experiments on Pascal-Part and self-collected in-the-wild animal datasets demonstrate considerably better 3D reconstructions as well as both 2D and 3D part discovery compared to prior arts. Project page: https://chhankyao.github.io/lassie/

Video salient object detection aims to find the most visually distinctive objects in a video. To explore the temporal dependencies, existing methods usually resort to recurrent neural networks or optical flow. However, these approaches require high computational cost, and tend to accumulate inaccuracies over time. In this paper, we propose a network with attention modules to learn contrastive features for video salient object detection without the high computational temporal modeling techniques. We develop a non-local self-attention scheme to capture the global information in the video frame. A co-attention formulation is utilized to combine the low-level and high-level features. We further apply the contrastive learning to improve the feature representations, where foreground region pairs from the same video are pulled together, and foreground-background region pairs are pushed away in the latent space. The intra-frame contrastive loss helps separate the foreground and background features, and the inter-frame contrastive loss improves the temporal consistency. We conduct extensive experiments on several benchmark datasets for video salient object detection and unsupervised video object segmentation, and show that the proposed method requires less computation, and performs favorably against the state-of-the-art approaches.

Multi-Task Learning (MTL) aims to enhance the model generalization by sharing representations between related tasks for better performance. Typical MTL methods are jointly trained with the complete multitude of ground-truths for all tasks simultaneously. However, one single dataset may not contain the annotations for each task of interest. To address this issue, we propose the Semi-supervised Multi-Task Learning (SemiMTL) method to leverage the available supervisory signals from different datasets, particularly for semantic segmentation and depth estimation tasks. To this end, we design an adversarial learning scheme in our semi-supervised training by leveraging unlabeled data to optimize all the task branches simultaneously and accomplish all tasks across datasets with partial annotations. We further present a domain-aware discriminator structure with various alignment formulations to mitigate the domain discrepancy issue among datasets. Finally, we demonstrate the effectiveness of the proposed method to learn across different datasets on challenging street view and remote sensing benchmarks.

Federated learning methods enable us to train machine learning models on distributed user data while preserving its privacy. However, it is not always feasible to obtain high-quality supervisory signals from users, especially for vision tasks. Unlike typical federated settings with labeled client data, we consider a more practical scenario where the distributed client data is unlabeled, and a centralized labeled dataset is available on the server. We further take the server-client and inter-client domain shifts into account and pose a domain adaptation problem with one source (centralized server data) and multiple targets (distributed client data). Within this new Federated Multi-Target Domain Adaptation (FMTDA) task, we analyze the model performance of existing domain adaptation methods and propose an effective DualAdapt method to address the new challenges. Extensive experimental results on image classification and semantic segmentation tasks demonstrate that our method achieves high accuracy, incurs minimal communication cost, and requires low computational resources on client devices.

Abstract Referring expression comprehension aims to localize objects identified by natural language descriptions. This is a challenging task as it requires understanding of both visual and language domains. One nature is that each object can be described by synonymous sentences with paraphrases, and such varieties in languages have critical impact on learning a comprehension model. While prior work usually treats each sentence and attends it to an object separately, we focus on learning a referring expression comprehension model that considers the property in synonymous sentences. To this end, we develop an end-to-end trainable framework to learn contrastive features on the image and object instance levels, where features extracted from synonymous sentences to describe the same object should be closer to each other after mapping to the visual domain. We conduct extensive experiments to evaluate the proposed algorithm on several benchmark datasets, and demonstrate that our method performs favorably against the state-of-the-art approaches. Furthermore, since the varieties in expressions become larger across datasets when they describe objects in different ways, we present the cross-dataset and transfer learning settings to validate the ability of our learned transferable features.

Joint filtering mainly uses an additional guidance image as a prior and transfers its structures to the target image in the filtering process. Different from existing approaches that rely on local linear models or hand-designed objective functions to extract the structural information from the guidance image, we propose a new joint filtering method based on a spatially variant linear representation model (SVLRM), where the target image is linearly represented by the guidance image. However, learning SVLRMs for vision tasks is a highly ill-posed problem. To estimate the spatially variant linear representation coefficients, we develop an effective approach based on a deep convolutional neural network (CNN). As such, the proposed deep CNN (constrained by the SVLRM) is able to model the structural information of both the guidance and input images. We show that the proposed approach can be effectively applied to a variety of applications, including depth/RGB image upsampling and restoration, flash deblurring, natural image denoising, and scale-aware filtering. In addition, we show that the linear representation model can be extended to high-order representation models (e.g., quadratic and cubic polynomial representations). Extensive experimental results demonstrate that the proposed method performs favorably against the state-of-the-art methods that have been specifically designed for each task.

Real-time semantic video segmentation is a challenging task due to the strict requirements of inference speed. Recent approaches mainly devote great efforts to reducing the model size for high efficiency. In this paper, we rethink this problem from a different viewpoint: using knowledge contained in compressed videos. We propose a simple and effective framework, dubbed TapLab, to tap into resources from the compressed domain. Specifically, we design a fast feature warping module using motion vectors for acceleration. To reduce the noise introduced by motion vectors, we design a residual-guided correction module and a residual-guided frame selection module using residuals. TapLab significantly reduces redundant computations of the state-of-the-art fast semantic image segmentation models, running 3 to 10 times faster with controllable accuracy degradation. The experimental results show that TapLab achieves 70.6 percent mIoU on the Cityscapes dataset at 99.8 FPS with a single GPU card for the 1024×2048 videos. A high-speed version even reaches the speed of 160+ FPS. Code will be available soon at https://github.com/Sixkplus/TapLab.

We present a learning-based approach for removing unwanted obstructions, such as window reflections, fence occlusions, or adherent raindrops, from a short sequence of images captured by a moving camera. Our method leverages motion differences between the background and obstructing elements to recover both layers. Specifically, we alternate between estimating dense optical flow fields of the two layers and reconstructing each layer from the flow-warped images via a deep convolutional neural network. This learning-based layer reconstruction module facilitates accommodating potential errors in the flow estimation and brittle assumptions, such as brightness consistency. We show that the proposed approach learned from synthetically generated data performs well to real images. Experimental results on numerous challenging scenarios of reflection and fence removal demonstrate the effectiveness of the proposed method.

In this paper, we address the issue of data imbalance in learning deep models for visual object tracking. Although it is well known that data distribution plays a crucial role in learning and inference models, considerably less attention has been paid to data imbalance in visual tracking. For the deep regression trackers that directly learn a dense mapping from input images of target objects to soft response maps, we identify their performance is limited by the extremely imbalanced pixel-to-pixel differences when computing regression loss. This prevents existing end-to-end learnable deep regression trackers from performing as well as discriminative correlation filters (DCFs) trackers. For the deep classification trackers that draw positive and negative samples to learn discriminative classifiers, there exists heavy class imbalance due to a limited number of positive samples when compared to the number of negative samples. To balance training data, we propose a novel shrinkage loss to penalize the importance of easy training data mostly coming from the background, which facilitates both deep regression and classification trackers to better distinguish target objects from the background. We extensively validate the proposed shrinkage loss function on six benchmark datasets, including the OTB-2013, OTB-2015, UAV-123, VOT-2016, VOT-2018 and LaSOT. Equipped with our shrinkage loss, the proposed one-stage deep regression tracker achieves favorable results against state-of-the-art methods, especially in comparison with DCFs trackers. Meanwhile, our shrinkage loss generalizes well to deep classification trackers. When replacing the original binary cross entropy loss with our shrinkage loss, three representative baseline trackers achieve large performance gains, even setting new state-of-the-art results.

Deblurring images captured in dynamic scenes is challenging as the motion blurs are spatially varying caused by camera shakes and object movements. In this paper, we propose a spatially varying neural network to deblur dynamic scenes. The proposed model is composed of three deep convolutional neural networks (CNNs) and a recurrent neural network (RNN). The RNN is used as a deconvolution operator on feature maps extracted from the input image by one of the CNNs. Another CNN is used to learn the spatially varying weights for the RNN. As a result, the RNN is spatial-aware and can implicitly model the deblurring process with spatially varying kernels. To better exploit properties of the spatially varying RNN, we develop both one-dimensional and two-dimensional RNNs for deblurring. The third component, based on a CNN, reconstructs the final deblurred feature maps into a restored image. In addition, the whole network is end-to-end trainable. Quantitative and qualitative evaluations on benchmark datasets demonstrate that the proposed method performs favorably against the state-of-the-art deblurring algorithms.

Most image super-resolution (SR) methods are developed on synthetic low-resolution (LR) and high-resolution (HR) image pairs that are constructed by a predetermined operation, e.g., bicubic downsampling. As existing methods typically learn an inverse mapping of the specific function, they produce blurry results when applied to real-world images whose exact formulation is different and unknown. Therefore, several methods attempt to synthesize much more diverse LR samples or learn a realistic downsampling model. However, due to restrictive assumptions on the downsampling process, they are still biased and less generalizable. This study proposes a novel method to simulate an unknown downsampling process without imposing restrictive prior knowledge. We propose a generalizable low-frequency loss (LFL) in the adversarial training framework to imitate the distribution of target LR images without using any paired examples. Furthermore, we design an adaptive data loss (ADL) for the downsampler, which can be adaptively learned and updated from the data during the training loops. Extensive experiments validate that our downsampling model can facilitate existing SR methods to perform more accurate reconstructions on various synthetic and real-world examples than the conventional approaches.

Super-resolution is a fundamental problem in computer vision which aims to overcome the spatial limitation of camera sensors. While significant progress has been made in single image super-resolution, most algorithms only perform well on synthetic data, which limits their applications in real scenarios. In this paper, we study the problem of real-scene single image super-resolution to bridge the gap between synthetic data and real captured images. We focus on two issues of existing super-resolution algorithms: lack of realistic training data and insufficient utilization of visual information obtained from cameras. To address the first issue, we propose a method to generate more realistic training data by mimicking the imaging process of digital cameras. For the second issue, we develop a two-branch convolutional neural network to exploit the radiance information originally-recorded in raw images. In addition, we propose a dense channel-attention block for better image restoration as well as a learning-based guided filter network for effective color correction. Our model is able to generalize to different cameras without deliberately training on images from specific camera types. Extensive experiments demonstrate that the proposed algorithm can recover fine details and clear structures, and achieve high-quality results for single image super-resolution in real scenes.

As an emerging and challenging problem in the computer vision community, weakly supervised object localization and detection plays an important role for developing new generation computer vision systems and has received significant attention in the past decade. As methods have been proposed, a comprehensive survey of these topics is of great importance. In this work, we review (1) classic models, (2) approaches with feature representations from off-the-shelf deep networks, (3) approaches solely based on deep learning, and (4) publicly available datasets and standard evaluation metrics that are widely used in this field. We also discuss the key challenges in this field, development history of this field, advantages/disadvantages of the methods in each category, the relationships between methods in different categories, applications of the weakly supervised object localization and detection methods, and potential future directions to further promote the development of this research field.

In this work, we propose a simple yet effective meta-learning algorithm in semi-supervised learning. We notice that most existing consistency-based approaches suffer from overfitting and limited model generalization ability, especially when training with only a small number of labeled data. To alleviate this issue, we propose a learn-to-generalize regularization term by utilizing the label information and optimize the problem in a meta-learning fashion. Specifically, we seek the pseudo labels of the unlabeled data so that the model can generalize well on the labeled data, which is formulated as a nested optimization problem. We address this problem using the meta-gradient that bridges between the pseudo label and the regularization term. In addition, we introduce a simple first-order approximation to avoid computing higher-order derivatives and provide theoretic convergence analysis. Extensive evaluations on the SVHN, CIFAR, and ImageNet datasets demonstrate that the proposed algorithm performs favorably against state-of-the-art methods.

Image-to-image (I2I) translation methods based on generative adversarial networks (GANs) typically suffer from overfitting when limited training data is available. In this work, we propose a data augmentation method (ReMix) to tackle this issue. We interpolate training samples at the feature level and propose a novel content loss based on the perceptual relations among samples. The generator learns to translate the in-between samples rather than memorizing the training set, and thereby forces the discriminator to generalize. The proposed approach effectively reduces the ambiguity of generation and renders content-preserving results. The ReMix method can be easily incorporated into existing GAN models with minor modifications. Experimental results on numerous tasks demonstrate that GAN models equipped with the ReMix method achieve significant improvements.

We present a self-supervised Contrastive Video Representation Learning (CVRL) method to learn spatiotemporal visual representations from unlabeled videos. Our representations are learned using a contrastive loss, where two augmented clips from the same short video are pulled together in the embedding space, while clips from different videos are pushed away. We study what makes for good data augmentations for video self-supervised learning and find that both spatial and temporal information are crucial. We carefully design data augmentations involving spatial and temporal cues. Concretely, we propose a temporally consistent spatial augmentation method to impose strong spatial augmentations on each frame of the video while maintaining the temporal consistency across frames. We also propose a sampling-based temporal augmentation method to avoid overly enforcing invariance on clips that are distant in time. On Kinetics-600, a linear classifier trained on the representations learned by CVRL achieves 70.4% top-1 accuracy with a 3D-ResNet-50 (R3D-50) backbone, outperforming ImageNet supervised pre-training by 15.7% and SimCLR unsupervised pre-training by 18.8% using the same inflated R3D-50. The performance of CVRL can be further improved to 72.9% with a larger R3D-152 (2× filters) backbone, significantly closing the gap between unsupervised and supervised video representation learning. Our code and models will be available at https://github.com/tensorflow/models/tree/master/official/.

Recent years have witnessed the rapid progress of generative adversarial networks (GANs). However, the success of the GAN models hinges on a large amount of training data. This work proposes a regularization approach for training robust GAN models on limited data. We theoretically show a connection between the regularized loss and an f-divergence called LeCam-divergence, which we find is more robust under limited training data. Extensive experiments on several benchmark datasets demonstrate that the proposed regularization scheme 1) improves the generalization performance and stabilizes the learning dynamics of GAN models under limited training data, and 2) complements the recent data augmentation methods. These properties facilitate training GAN models to achieve state-of-theart performance when only limited training data of the ImageNet benchmark is available. The source code is available at https://github.com/google/lecam-gan.

Image restoration tasks demand a complex balance between spatial details and high-level contextualized information while recovering images. In this paper, we propose a novel synergistic design that can optimally balance these competing goals. Our main proposal is a multi-stage architecture, that progressively learns restoration functions for the degraded inputs, thereby breaking down the overall recovery process into more manageable steps. Specifically, our model first learns the contextualized features using encoder-decoder architectures and later combines them with a high-resolution branch that retains local information. At each stage, we introduce a novel per-pixel adaptive design that leverages in-situ supervised attention to reweight the local features. A key ingredient in such a multi-stage architecture is the information exchange between different stages. To this end, we propose a two-faceted approach where the information is not only exchanged sequentially from early to late stages, but lateral connections between feature processing blocks also exist to avoid any loss of information. The resulting tightly interlinked multi-stage architecture, named as MPRNet, delivers strong performance gains on ten datasets across a range of tasks including image deraining, deblurring, and denoising. The source code and pre-trained models are available at https://github.com/swz30/MPRNet.

Convolution is one of the basic building blocks of CNN architectures. Despite its common use, standard convolution has two main shortcomings: Content-agnostic and Computation-heavy. Dynamic filters are content-adaptive, while further increasing the computational overhead. Depth-wise convolution is a lightweight variant, but it usually leads to a drop in CNN performance or requires a larger number of channels. In this work, we propose the Decoupled Dynamic Filter (DDF) that can simultaneously tackle both of these shortcomings. Inspired by recent advances in attention, DDF decouples a depth-wise dynamic filter into spatial and channel dynamic filters. This decomposition considerably reduces the number of parameters and limits computational costs to the same level as depth-wise convolution. Meanwhile, we observe a significant boost in performance when replacing standard convolution with DDF in classification networks. ResNet50 / 101 get improved by 1.9% and 1.3% on the top-1 accuracy, while their computational costs are reduced by nearly half. Experiments on the detection and joint upsampling networks also demonstrate the superior performance of the DDF upsampling variant (DDF-Up) in comparison with standard convolution and specialized content-adaptive layers. The project page with code is available1.

We tackle a 3D scene stylization problem — generating stylized images of a scene from arbitrary novel views given a set of images of the same scene and a reference image of the desired style as inputs. Direct solution of combining novel view synthesis and stylization approaches lead to results that are blurry or not consistent across different views. We propose a point cloud-based method for consistent 3D scene stylization. First, we construct the point cloud by back-projecting the image features to the 3D space. Second, we develop point cloud aggregation modules to gather the style information of the 3D scene, and then modulate the features in the point cloud with a linear transformation matrix. Finally, we project the transformed features to 2D space to obtain the novel views. Experimental results on two diverse datasets of real-world scenes validate that our method generates consistent stylized novel view synthesis results against other alternative approaches.

Most video super-resolution methods focus on restoring high-resolution video frames from low-resolution videos without taking into account compression. However, most videos on the web or mobile devices are compressed, and the compression can be severe when the bandwidth is limited. In this paper, we propose a new compression-informed video super-resolution model to restore high-resolution content without introducing artifacts caused by compression. The proposed model consists of three modules for video super-resolution: bi-directional recurrent warping, detail-preserving flow estimation, and Laplacian enhancement. All these three modules are used to deal with compression properties such as the location of the intra-frames in the input and smoothness in the output frames. For thorough performance evaluation, we conducted extensive experiments on standard datasets with a wide range of compression rates, covering many real video use cases. We showed that our method not only recovers high-resolution content on uncompressed frames from the widely-used benchmark datasets, but also achieves state-of-the-art performance in super-resolving compressed videos based on numerous quantitative metrics. We also evaluated the proposed method by simulating streaming from YouTube to demonstrate its effectiveness and robustness. The source codes and trained models are available at https://github.com/google-research/googleresearch/tree/master/comisr.

Existing video stabilization methods often generate visible distortion or require aggressive cropping of frame boundaries, resulting in smaller field of views. In this work, we present a frame synthesis algorithm to achieve full-frame video stabilization. We first estimate dense warp fields from neighboring frames and then synthesize the stabilized frame by fusing the warped contents. Our core technical novelty lies in the learning-based hybrid-space fusion that alleviates artifacts caused by optical flow inaccuracy and fast-moving objects. We validate the effectiveness of our method on the NUS, selfie, and DeepStab video datasets. Extensive experiment results demonstrate the merits of our approach over prior video stabilization methods.

This work proposes a weakly-supervised temporal action localization framework, called D2-Net, which strives to temporally localize actions using video-level supervision. Our main contribution is the introduction of a novel loss formulation, which jointly enhances the discriminability of latent embeddings and robustness of the output temporal class activations with respect to foreground-background noise caused by weak supervision. The proposed formulation comprises a discriminative and a denoising loss term for enhancing temporal action localization. The discriminative term incorporates a classification loss and utilizes a top-down attention mechanism to enhance the separability of latent foreground-background embeddings. The denoising loss term explicitly addresses the foreground-background noise in class activations by simultaneously maximizing intra-video and inter-video mutual information using a bottom-up attention mechanism. As a result, activations in the foreground regions are emphasized whereas those in the background regions are suppressed, thereby leading to more robust predictions. Comprehensive experiments are performed on multiple benchmarks, including THUMOS14 and ActivityNet1.2. Our D2-Net performs favorably in comparison to the existing methods on all datasets, achieving gains as high as 2.3% in terms of mAP at IoU=0.5 on THUMOS14. Source code is available at https://github.com/naraysa/D2-Net.

Vision-and-Language Navigation (VLN) requires an agent to find a path to a remote location on the basis of natural-language instructions and a set of photo-realistic panoramas. Most existing methods take the words in the instructions and the discrete views of each panorama as the minimal unit of encoding. However, this requires a model to match different nouns (e.g., TV, table) against the same input view feature. In this work, we propose an object-informed sequential BERT to encode visual perceptions and linguistic instructions at the same fine-grained level, namely objects and words. Our sequential BERT also enables the visual-textual clues to be interpreted in light of the temporal context, which is crucial to multi-round VLN tasks. Additionally, we enable the model to identify the relative direction (e.g., left/right/front/back) of each navigable location and the room type (e.g., bedroom, kitchen) of its current and final navigation goal, as such information is widely mentioned in instructions implying the desired next and final locations. We thus enable the model to know-where the objects lie in the images, and to know-where they stand in the scene. Extensive experiments demonstrate the effectiveness compared against several state-of-the-art methods on three indoor VLN tasks: REVERIE, NDH, and R2R. Project repository: https://github.com/YuankaiQi/ORIST

Learning temporally consistent foreground opacity from videos, i.e., video matting, has drawn great attention due to the blossoming of video conferencing. Previous approaches are built on top of image matting models, which fail in maintaining the temporal coherence when being adapted to videos. They either utilize the optical flow to smooth frame-wise prediction, where the performance is dependent on the selected optical flow model; or naively combine feature maps from multiple frames, which does not model well the correspondence of pixels in adjacent frames. In this paper, we propose to enhance the temporal coherence by Consistency-Regularized Graph Neural Networks (CRGNN) with the aid of a synthesized video matting dataset. CRGNN utilizes Graph Neural Networks (GNN) to relate adjacent frames such that pixels or regions that are incorrectly predicted in one frame can be corrected by leveraging information from its neighboring frames. To generalize our model from synthesized videos to real-world videos, we propose a consistency regularization technique to enforce the consistency on the alpha and foreground when blending them with different backgrounds. To evaluate the efficacy of CRGNN, we further collect a real-world dataset with annotated alpha mattes. Compared with state-of-the-art methods that require hand-crafted trimaps or backgrounds for modeling training, CRGNN generates favorably results with the help of unlabeled real training dataset. The source code and datasets are available at https://github.com/TiantianWang/VideoMattingCRGNN.git.

Reasoning 3D shapes from 2D images is an essential yet challenging task, especially when only single-view images are at our disposal. While an object can have a complicated shape, individual parts are usually close to geometric primitives and thus are easier to model. Furthermore, parts provide a mid-level representation that is robust to appearance variations across objects in a particular category. In this work, we tackle the problem of 3D part discovery from only 2D image collections. Instead of relying on manually annotated parts for supervision, we propose a self-supervised approach, latent part discovery (LPD). Our key insight is to learn a novel part shape prior that allows each part to fit an object shape faithfully while constrained to have simple geometry. Extensive experiments on the synthetic ShapeNet, PartNet, and real-world Pascal 3D+ datasets show that our method discovers consistent object parts and achieves favorable reconstruction accuracy compared to the existing methods with the same level of supervision. Our project page with code is at https://chhankyao.github.io/lpd/.

Increasingly, modern mobile devices allow capturing images at Ultra-High-Definition (UHD) resolution, which includes 4K and 8K images. However, current single image super-resolution (SISR) methods focus on super-resolving images to ones with resolution up to high definition (HD) and ignore higher-resolution UHD images. To explore their performance on UHD images, in this paper, we first introduce two large-scale image datasets, UHDSR4K and UHDSR8K, to benchmark existing SISR methods. With 70,000 V100 GPU hours of training, we benchmark these methods on 4K and 8K resolution images under seven different settings to provide a set of baseline models. Moreover, we propose a baseline model, called Mesh Attention Network (MANet) for SISR. The MANet applies the attention mechanism in both different depths (horizontal) and different levels of receptive field (vertical). In this way, correlations among feature maps are learned, enabling the network to focus on more important features.

We address the problem of text-guided video temporal grounding, which aims to identify the time interval of a certain event based on a natural language description. Different from most existing methods that only consider RGB images as visual features, we propose a multi-modal framework to extract complementary information from videos. Specifically, we adopt RGB images for appearance, optical flow for motion, and depth maps for image structure. While RGB images provide abundant visual cues of certain events, the performance may be affected by background clutters. Therefore, we use optical flow to focus on large motion and depth maps to infer the scene configuration when the action is related to objects recognizable with their shapes. To integrate the three modalities more effectively and enable inter-modal learning, we design a dynamic fusion scheme with transformers to model the interactions between modalities. Furthermore, we apply intra-modal self-supervised learning to enhance feature representations across videos for each modality, which also facilitates multi-modal learning. We conduct extensive experiments on the Charades-STA and ActivityNet Captions datasets, and show that the proposed method performs favorably against state-of-the-art approaches.

We propose a canonical point autoencoder (CPAE) that predicts dense correspondences between 3D shapes of the same category. The autoencoder performs two key functions: (a) encoding an arbitrarily ordered point cloud to a canonical primitive, e.g., a sphere, and (b) decoding the primitive back to the original input instance shape. As being placed in the bottleneck, this primitive plays a key role to map all the unordered point clouds on the canonical surface, and to be reconstructed in an ordered fashion. Once trained, points from different shape instances that are mapped to the same locations on the primitive surface are determined to be a pair of correspondence. Our method does not require any form of annotation or self-supervised part segmentation network and can handle unaligned input point clouds within a certain rotation range. Experimental results on 3D semantic keypoint transfer and part segmentation transfer show that our model performs favorably against state-of-the-art correspondence learning methods.

The audio-visual video parsing task aims to temporally parse a video into audio or visual event categories. However, it is labor intensive to temporally annotate audio and visual events and thus hampers the learning of a parsing model. To this end, we propose to explore additional cross-video and cross-modality supervisory signals to facilitate weakly-supervised audio-visual video parsing. The proposed method exploits both the common and diverse event semantics across videos to identify audio or visual events. In addition, our method explores event co-occurrence across audio, visual, and audio-visual streams. We leverage the explored cross-modality co-occurrence to localize segments of target events while excluding irrelevant ones. The discovered supervisory signals across different videos and modalities can greatly facilitate the training with only video-level annotations. Quantitative and qualitative results demonstrate that the proposed method performs favorably against existing methods on weakly-supervised audio-visual video parsing.

Vision transformers (ViT) have demonstrated impressive performance across numerous machine vision tasks. These models are based on multi-head self-attention mechanisms that can flexibly attend to a sequence of image patches to encode contextual cues. An important question is how such flexibility (in attending image-wide context conditioned on a given patch) can facilitate handling nuisances in natural images e.g., severe occlusions, domain shifts, spatial permutations, adversarial and natural perturbations. We systematically study this question via an extensive set of experiments encompassing three ViT families and provide comparisons with a high-performing convolutional neural network (CNN). We show and analyze the following intriguing properties of ViT: (a)Transformers are highly robust to severe occlusions, perturbations and domain shifts, e.g., retain as high as 60% top-1 accuracy on ImageNet even after randomly occluding 80% of the image content. (b)The robustness towards occlusions is not due to texture bias, instead we show that ViTs are significantly less biased towards local textures, compared to CNNs. When properly trained to encode shape-based features, ViTs demonstrate shape recognition capability comparable to that of human visual system, previously unmatched in the literature. (c)Using ViTs to encode shape representation leads to an interesting consequence of accurate semantic segmentation without pixel-level supervision. (d)Off-the-shelf features from a single ViT model can be combined to create a feature ensemble, leading to high accuracy rates across a range of classification datasets in both traditional and few-shot learning paradigms. We show effective features of ViTs are due to flexible and dynamic receptive fields possible via self-attention mechanisms. Our code will be publicly released.

Recent advances in convolutional neural networks (CNNs) usually come with the expense of excessive computational overhead and memory footprint. Network compression aims to alleviate this issue by training compact models with comparable performance. However, existing compression techniques either entail dedicated expert design or compromise with a moderate performance drop. In this paper, we propose a novel structured sparsification method for efficient network compression. The proposed method automatically induces structured sparsity on the convolutional weights, thereby facilitating the implementation of the compressed model with the highly-optimized group convolution. We further address the problem of inter-group communication with a learnable channel shuffle mechanism. The proposed approach can be easily applied to compress many network architectures with a negligible performance drop. Extensive experimental results and analysis demonstrate that our approach gives a competitive performance against the recent network compression counterparts with a sound accuracy-complexity trade-off.

Image extrapolation aims at expanding the narrow field of view of a given image patch. Existing models mainly deal with natural scene images of homogeneous regions and have no control of the content generation process. In this work, we study conditional image extrapolation to synthesize new images guided by the input structured text. The text is represented as a graph to specify the objects and their spatial relation to the unknown regions of the image. Inspired by drawing techniques, we propose a progressive generative model of three stages, i.e., generating a coarse bounding-boxes layout, refining it to a finer segmentation layout, and mapping the layout to a realistic output. Such a multi-stage design is shown to facilitate the training process and generate more controllable results. We validate the effectiveness of the proposed method on the face and human clothing dataset in terms of visual results, quantitative evaluations, and flexible controls.

Learning multiple domains/tasks with a single model is important for improving data efficiency and lowering inference cost for numerous vision tasks, especially on resource-constrained mobile devices. However, hand-crafting a multi-domain/task model can be both tedious and challenging. This paper proposes a novel approach to automatically learn a multi-path network for multi-domain visual classification on mobile devices. The proposed multi-path network is learned from neural architecture search by applying one reinforcement learning controller for each domain to select the best path in the super-network created from a MobileNetV3-like search space. An adaptive balanced domain prioritization algorithm is proposed to balance optimizing the joint model on multiple domains simultaneously. The determined multi-path model selectively shares parameters across domains in shared nodes while keeping domain-specific parameters within non-shared nodes in individual domain paths. This approach effectively reduces the total number of parameters and FLOPS, encouraging positive knowledge transfer while mitigating negative interference across domains. Extensive evaluations on the Visual Decathlon dataset demonstrate that the proposed multi-path model achieves state-of-the-art performance in terms of accuracy, model size, and FLOPS against other approaches using MobileNetV3-like architectures. Furthermore, the proposed method improves average accuracy over learning single-domain models individually, and reduces the total number of parameters and FLOPS by 78% and 32% respectively, compared to the approach that simply bundles single-domain models for multi-domain learning.

Abstract Caricature is an artistic drawing created to abstract or exaggerate facial features of a person. Rendering visually pleasing caricatures is a difficult task that requires professional skills, and thus it is of great interest to design a method to automatically generate such drawings. To deal with large shape changes, we propose an algorithm based on a semantic shape transform to produce diverse and plausible shape exaggerations. Specifically, we predict pixel-wise semantic correspondences and perform image warping on the input photo to achieve dense shape transformation. We show that the proposed framework is able to render visually pleasing shape exaggerations while maintaining their facial structures. In addition, our model allows users to manipulate the shape via the semantic map. We demonstrate the effectiveness of our approach on a large photograph-caricature benchmark dataset with comparisons to the state-of-the-art methods.

Motion estimation (ME) and motion compensation (MC) have been widely used for classical video frame interpolation systems over the past decades. Recently, a number of data-driven frame interpolation methods based on convolutional neural networks have been proposed. However, existing learning based methods typically estimate either flow or compensation kernels, thereby limiting performance on both computational efficiency and interpolation accuracy. In this work, we propose a motion estimation and compensation driven neural network for video frame interpolation. A novel adaptive warping layer is developed to integrate both optical flow and interpolation kernels to synthesize target frame pixels. This layer is fully differentiable such that both the flow and kernel estimation networks can be optimized jointly. The proposed model benefits from the advantages of motion estimation and compensation methods without using hand-crafted features. Compared to existing methods, our approach is computationally efficient and able to generate more visually appealing results. Furthermore, the proposed MEMC-Net architecture can be seamlessly adapted to several video enhancement tasks, e.g., super-resolution, denoising, and deblocking. Extensive quantitative and qualitative evaluations demonstrate that the proposed method performs favorably against the state-of-the-art video frame interpolation and enhancement algorithms on a wide range of datasets.

We present an approach for jointly matching and segmenting object instances of the same category within a collection of images. In contrast to existing algorithms that tackle the tasks of semantic matching and object co-segmentation in isolation, our method exploits the complementary nature of the two tasks. The key insights of our method are two-fold. First, the estimated dense correspondence fields from semantic matching provide supervision for object co-segmentation by enforcing consistency between the predicted masks from a pair of images. Second, the predicted object masks from object co-segmentation in turn allow us to reduce the adverse effects due to background clutters for improving semantic matching. Our model is end-to-end trainable and does not require supervision from manually annotated correspondences and object masks. We validate the efficacy of our approach on five benchmark datasets: TSS, Internet, PF-PASCAL, PF-WILLOW, and SPair-71k, and show that our algorithm performs favorably against the state-of-the-art methods on both semantic matching and object co-segmentation tasks.

Representing features at multiple scales is of great importance for numerous vision tasks. Recent advances in backbone convolutional neural networks (CNNs) continually demonstrate stronger multi-scale representation ability, leading to consistent performance gains on a wide range of applications. However, most existing methods represent the multi-scale features in a layer-wise manner. In this paper, we propose a novel building block for CNNs, namely Res2Net, by constructing hierarchical residual-like connections within one single residual block. The Res2Net represents multi-scale features at a granular level and increases the range of receptive fields for each network layer. The proposed Res2Net block can be plugged into the state-of-the-art backbone CNN models, e.g., ResNet, ResNeXt, and DLA. We evaluate the Res2Net block on all these models and demonstrate consistent performance gains over baseline models on widely-used datasets, e.g., CIFAR-100 and ImageNet. Further ablation studies and experimental results on representative computer vision tasks, i.e., object detection, class activation mapping, and salient object detection, further verify the superiority of the Res2Net over the state-of-the-art baseline methods. The source code and trained models are available on https://mmcheng.net/res2net/.

We present an algorithm to directly solve numerous image restoration problems (e.g., image deblurring, image dehazing, and image deraining). These problems are ill-posed, and the common assumptions for existing methods are usually based on heuristic image priors. In this paper, we show that these problems can be solved by generative models with adversarial learning. However, a straightforward formulation based on a straightforward generative adversarial network (GAN) does not perform well in these tasks, and some structures of the estimated images are usually not preserved well. Motivated by an interesting observation that the estimated results should be consistent with the observed inputs under the physics models, we propose an algorithm that guides the estimation process of a specific task within the GAN framework. The proposed model is trained in an end-to-end fashion and can be applied to a variety of image restoration and low-level vision problems. Extensive experiments demonstrate that the proposed method performs favorably against state-of-the-art algorithms.

Most existing non-blind restoration methods are based on the assumption that a precise degradation model is known. As the degradation process can only be partially known or inaccurately modeled, images may not be well restored. Rain streak removal and image deconvolution with inaccurate blur kernels are two representative examples of such tasks. For rain streak removal, although an input image can be decomposed into a scene layer and a rain streak layer, there exists no explicit formulation for modeling rain streaks and the composition with scene layer. For blind deconvolution, as estimation error of blur kernel is usually introduced, the subsequent non-blind deconvolution process does not restore the latent image well. In this paper, we propose a principled algorithm within the maximum a posterior framework to tackle image restoration with a partially known or inaccurate degradation model. Specifically, the residual caused by a partially known or inaccurate degradation model is spatially dependent and complexly distributed. With a training set of degraded and ground-truth image pairs, we parameterize and learn the fidelity term for a degradation model in a task-driven manner. Furthermore, the regularization term can also be learned along with the fidelity term, thereby forming a simultaneous fidelity and regularization learning model. Extensive experimental results demonstrate the effectiveness of the proposed model for image deconvolution with inaccurate blur kernels, deconvolution with multiple degradations and rain streak removal.

Visual events are usually accompanied by sounds in our daily lives. However, can the machines learn to correlate the visual scene and sound, as well as localize the sound source only by observing them like humans? To investigate its empirical learnability, in this work we first present a novel unsupervised algorithm to address the problem of localizing sound sources in visual scenes. In order to achieve this goal, a two-stream network structure which handles each modality with attention mechanism is developed for sound source localization. The network naturally reveals the localized response in the scene without human annotation. In addition, a new sound source dataset is developed for performance evaluation. Nevertheless, our empirical evaluation shows that the unsupervised method generates false conclusions in some cases. Thereby, we show that this false conclusion cannot be fixed without human prior knowledge due to the well-known correlation and causality mismatch misconception. To fix this issue, we extend our network to the supervised and semi-supervised network settings via a simple modification due to the general architecture of our two-stream network. We show that the false conclusions can be effectively corrected even with a small amount of supervision, i.e., semi-supervised setup. Furthermore, we present the versatility of the learned audio and visual embeddings on the cross-modal content alignment and we extend this proposed algorithm to a new application, sound saliency based automatic camera view panning in 360 degree videos.

Facilitated by deep neural networks, numerous tracking methods have made significant advances. Existing deep trackers mainly utilize independent frames to model the target appearance, while paying less attention to its temporal coherence. In this paper, we propose a recurrent memory activation network (RMAN) to exploit the untapped temporal coherence of the target appearance for visual tracking. We build the RMAN on top of the long short-term memory network (LSTM) with an additional memory activation layer. Specifically, we first use the LSTM to model the temporal changes of the target appearance. Then we selectively activate the memory blocks via the activation layer to produce a temporally coherent representation. The recurrent memory activation layer enriches the target representations from independent frames and reduces the background interference through temporal consistency. The proposed RMAN is fully differentiable and can be optimized end-to-end. To facilitate network training, we propose a temporal coherence loss together with the original binary classification loss. Extensive experimental results on standard benchmarks demonstrate that our method performs favorably against the state-of-the-art approaches.

Pedestrian detection is a crucial task in intelligent transportation systems, which can be applied in autonomous vehicles and traffic scene video surveillance systems. The past few years have witnessed much progress on the research of pedestrian detection methods, especially through the successful use of the deep learning based techniques. However, occlusion and large scale variation remain the challenging issues for pedestrian detection. In this work, we propose a Part-Aware Multi-Scale Fully Convolutional Network (PAMS-FCN) to tackle these difficulties. Specifically, we present a part-aware Region-of-Interest (RoI) pooling module to mine body parts with different responses, and select the part with the strongest response via voting. As such, a partially visible pedestrian instance can receive a high detection confidence score, making it less likely to become a missing detection. This module operates in parallel with an instance RoI pooling module to combine local parts and global context information. To handle vast scale variation, we construct a fully convolutional network in which multi-scale feature maps are generated efficiently, and small-scale and large-scale pedestrians are detected separately. By integrating these structures, the proposed detector achieves the state-of-the-art performance on the Caltech, KITTI, INRIA and ETH pedestrian detection datasets.

Data privacy has emerged as an important issue as data-driven deep learning has been an essential component of modern machine learning systems. For instance, there could be a potential privacy risk of machine learning systems via the model inversion attack, whose goal is to reconstruct the input data from the latent representation of deep networks. Our work aims at learning a privacy-preserving and task-oriented representation to defend against such model inversion attacks. Specifically, we propose an adversarial reconstruction learning framework that prevents the latent representations decoded into original input data. By simulating the expected behavior of adversary, our framework is realized by minimizing the negative pixel reconstruction loss or the negative feature reconstruction (i.e., perceptual distance) loss. We validate the proposed method on face attribute prediction, showing that our method allows protecting visual privacy with a small decrease in utility performance. In addition, we show the utility-privacy trade-off with different choices of hyperparameter for negative perceptual distance loss at training, allowing service providers to determine the right level of privacy-protection with a certain utility performance. Moreover, we provide an extensive study with different selections of features, tasks, and the data to further analyze their influence on privacy protection.

Existing weakly-supervised semantic segmentation methods using image-level annotations typically rely on initial responses to locate object regions. However, such response maps generated by the classification network usually focus on discriminative object parts, due to the fact that the network does not need the entire object for optimizing the objective function. To enforce the network to pay attention to other parts of an object, we propose a simple yet effective approach that introduces a self-supervised task by exploiting the sub-category information. Specifically, we perform clustering on image features to generate pseudo sub-categories labels within each annotated parent class, and construct a sub-category objective to assign the network to a more challenging task. By iteratively clustering image features, the training process does not limit itself to the most discriminative object parts, hence improving the quality of the response maps. We conduct extensive analysis to validate the proposed method and show that our approach performs favorably against the state-of-the-art approaches.

In this paper, we propose a Multi-Scale Boosted Dehazing Network with Dense Feature Fusion based on the U-Net architecture. The proposed method is designed based on two principles, boosting and error feedback, and we show that they are suitable for the dehazing problem. By incorporating the Strengthen-Operate-Subtract boosting strategy in the decoder of the proposed model, we develop a simple yet effective boosted decoder to progressively restore the haze-free image. To address the issue of preserving spatial information in the U-Net architecture, we design a dense feature fusion module using the back-projection feedback scheme. We show that the dense feature fusion module can simultaneously remedy the missing spatial information from high-resolution features and exploit the non-adjacent features. Extensive evaluations demonstrate that the proposed model performs favorably against the state-of-the-art approaches on the benchmark datasets as well as real-world hazy images.

Object frequency in the real world often follows a power law, leading to a mismatch between datasets with long-tailed class distributions seen by a machine learning model and our expectation of the model to perform well on all classes. We analyze this mismatch from a domain adaptation point of view. First of all, we connect existing class-balanced methods for long-tailed classification to target shift, a well-studied scenario in domain adaptation. The connection reveals that these methods implicitly assume that the training data and test data share the same class-conditioned distribution, which does not hold in general and especially for the tail classes. While a head class could contain abundant and diverse training examples that well represent the expected data at inference time, the tail classes are often short of representative training data. To this end, we propose to augment the classic class-balanced learning by explicitly estimating the differences between the class-conditioned distributions with a meta-learning approach. We validate our approach with six benchmark datasets and three loss functions.

Finding views with a good composition from an input image is a common but challenging problem. There are usually at least dozens of candidates (regions) in an image, and how to evaluate these candidates is subjective. Most existing methods only use the feature corresponding to each candidate to evaluate the quality. However, the mutual relations between the candidates from an image play an essential role in composing a good shot due to the comparative nature of this problem. Motivated by this, we propose a graph-based module with a gated feature update to model the relations between different candidates. The candidate region features are propagated on a graph that models mutual relations between different regions for mining the useful information such that the relation features and region features are adaptively fused. We design a multi-task loss to train the model, especially, a regularization term is adopted to incorporate the prior knowledge about the relations into the graph. A data augmentation method is also developed by mixing nodes from different graphs to improve the model generalization ability. Experimental results show that the proposed model performs favorably against state-of-the-art methods, and comprehensive ablation studies demonstrate the contribution of each module and graph-based inference of the proposed method.

We present a learning-based approach for removing unwanted obstructions, such as window reflections, fence occlusions or raindrops, from a short sequence of images captured by a moving camera. Our method leverages the motion differences between the background and the obstructing elements to recover both layers. Specifically, we alternate between estimating dense optical flow fields of the two layers and reconstructing each layer from the flow-warped images via a deep convolutional neural network. The learning-based layer reconstruction allows us to accommodate potential errors in the flow estimation and brittle assumptions such as brightness consistency. We show that training on synthetically generated data transfers well to real images. Our results on numerous challenging scenarios of reflection and fence removal demonstrate the effectiveness of the proposed method.

Recovering a high dynamic range (HDR) image from a single low dynamic range (LDR) input image is challenging due to missing details in under-/over-exposed regions caused by quantization and saturation of camera sensors. In contrast to existing learning-based methods, our core idea is to incorporate the domain knowledge of the LDR image formation pipeline into our model. We model the HDR-to-LDR image formation pipeline as the (1) dynamic range clipping, (2) non-linear mapping from a camera response function, and (3) quantization. We then propose to learn three specialized CNNs to reverse these steps. By decomposing the problem into specific sub-tasks, we impose effective physical constraints to facilitate the training of individual sub-networks. Finally, we jointly fine-tune the entire model end-to-end to reduce error accumulation. With extensive quantitative and qualitative experiments on diverse image datasets, we demonstrate that the proposed method performs favorably against state-of-the-art single-image HDR reconstruction algorithms.

Universal style transfer methods typically leverage rich representations from deep Convolutional Neural Network (CNN) models (e.g., VGG-19) pre-trained on large collections of images. Despite the effectiveness, its application is heavily constrained by the large model size to handle ultra-resolution images given limited memory. In this work, we present a new knowledge distillation method (named Collaborative Distillation) for encoder-decoder based neural style transfer to reduce the convolutional filters. The main idea is underpinned by a finding that the encoder-decoder pairs construct an exclusive collaborative relationship, which is regarded as a new kind of knowledge for style transfer models. Moreover, to overcome the feature size mismatch when applying collaborative distillation, a linear embedding loss is introduced to drive the student network to learn a linear embedding of the teacher’s features. Extensive experiments show the effectiveness of our method when applied to different universal style transfer approaches (WCT and AdaIN), even if the model size is reduced by 15.5 times. Especially, on WCT with the compressed models, we achieve ultra-resolution (over 40 megapixels) universal style transfer on a 12GB GPU for the first time. Further experiments on optimization-based stylization scheme show the generality of our algorithm on different stylization paradigms. Our code and trained models are available at https://github.com/mingsun-tse/collaborative-distillation.

The availability of large-scale datasets has helped unleash the true potential of deep convolutional neural networks (CNNs). However, for the single-image denoising problem, capturing a real dataset is an unacceptably expensive and cumbersome procedure. Consequently, image denoising algorithms are mostly developed and evaluated on synthetic data that is usually generated with a widespread assumption of additive white Gaussian noise (AWGN). While the CNNs achieve impressive results on these synthetic datasets, they do not perform well when applied on real camera images, as reported in recent benchmark datasets. This is mainly because the AWGN is not adequate for modeling the real camera noise which is signal-dependent and heavily transformed by the camera imaging pipeline. In this paper, we present a framework that models camera imaging pipeline in forward and reverse directions. It allows us to produce any number of realistic image pairs for denoising both in RAW and sRGB spaces. By training a new image denoising network on realistic synthetic data, we achieve the state-of-the-art performance on real camera benchmark datasets. The parameters in our models are ~5 times lesser than the previous best method for RAW denoising. Furthermore, we demonstrate that the proposed framework generalizes beyond image denoising problem e.g., for color matching in stereoscopic cinema. The source code and pre-trained models are available at https://github.com/swz30/CycleISP.

We propose a new deep convolutional neural network framework that uses object location knowledge implicit in network connection weights to guide selective attention in object detection tasks. Our approach is called What-Where Nets (WW-Nets), and it is inspired by the structure of human visual pathways. In the brain, vision incorporates two separate streams, one in the temporal lobe and the other in the parietal lobe, called the ventral stream and the dorsal stream, respectively. The ventral pathway from primary visual cortex is dominated by "what" information, while the dorsal pathway is dominated by "where" information. Inspired by this structure, we have proposed an object detection framework involving the integration of a "What Network" and a "Where Network". The aim of the What Network is to provide selective attention to the relevant parts of the input image. The Where Network uses this information to locate and classify objects of interest. In this paper, we compare this approach to state-of-the-art algorithms on the PASCAL VOC 2007 and 2012 and COCO object detection challenge datasets. Also, we compare out approach to human "ground-truth" attention. We report the results of an eye-tracking experiment on human subjects using images from PASCAL VOC 2007, and we demonstrate interesting relationships between human overt attention and information processing in our WW-Nets. Finally, we provide evidence that our proposed method performs favorably in comparison to other object detection approaches, often by a large margin. The code and the eye-tracking ground-truth dataset can be found at: https://github.com/mkebrahimpour.

This paper presents an algorithm to reconstruct temporally consistent 3D meshes of deformable object instances from videos in the wild. Without requiring annotations of 3D mesh, 2D keypoints, or camera pose for each video frame, we pose video-based reconstruction as a self-supervised online adaptation problem applied to any incoming test video. We first learn a category-specific 3D reconstruction model from a collection of single-view images of the same category that jointly predicts the shape, texture, and camera pose of an image. Then, at inference time, we adapt the model to a test video over time using self-supervised regularization terms that exploit temporal consistency of an object instance to enforce that all reconstructed meshes share a common texture map, a base shape, as well as parts. We demonstrate that our algorithm recovers temporally consistent and reliable 3D structures from videos of non-rigid objects including those of animals captured in the wild -- an extremely challenging task rarely addressed before.

In this work, we introduce VQA 360°, a novel task of visual question answering on 360° images. Unlike a normal field-of-view image, a 360° image captures the entire visual content around the optical center of a camera, demanding more sophisticated spatial understanding and reasoning. To address this problem, we collect the first VQA 360° dataset, containing around 17,000 real-world image-question-answer triplets for a variety of question types. We then study two different VQA models on VQA 360°, including one conventional model that takes an equirectangular image (with intrinsic distortion) as input and one dedicated model that first projects a 360° image onto cubemaps and subsequently aggregates the information from multiple spatial resolutions. We demonstrate that the cubemap-based model with multi-level fusion and attention diffusion performs favorably against other variants and the equirectangular-based models. Nevertheless, the gap between the humans' and machines' performance reveals the need for more advanced VQA 360° algorithms. We, therefore, expect our dataset and studies to serve as the benchmark for future development in this challenging task. Dataset, code, and pre-trained models are available online.

Recent progress in video style transfer has shown promising results which contain less flickering effects. However, existing algorithms mainly trade off generality for efficiency, i.e., constructing one network per style example, and often work for short video clips only. In this work, we propose a video multi-style transfer (VMST) framework which enables fast and multi-style video transfer within one single network. Specifically, we design a multi-instance normalization block (MIN-Block) to learn different style examples and two ConvLSTM modules to encourage the temporal consistency. The proposed algorithm is demonstrated to be able to generate temporally-consistent video transfer results in different styles while keeping each stylized frame visually pleasing. Extensive experimental results show that the proposed method performs favorably against single-style models and some post-processing techniques that alleviate the flickering issue. We achieve as many as 120 stylization effects in a single model and show results on long-term videos that consist of thousands of frames.

Hashing has attracted attention in recent years due to the rapid growth of image and video data on the web. Benefiting from recent advances in deep learning, deep supervised hashing has achieved promising results for image retrieval. However, existing methods are either less efficient in data usage or incapable of learning linearly discriminative binary codes. In this paper, we revisit linear discriminative analysis and propose a linear discriminative hashing (LDH) objective that is efficient in training and achieves better accuracy in retrieval. With the joint supervision of a classification loss, we design a robust deep network to obtain binary codes that are inter-class separable and intra-class compact, which provides better representations for image retrieval. We conduct extensive experiments on three benchmark datasets, and our LDH algorithm performs favorably against existing state-of-the-art deep supervised hashing methods.

Recent deep learning methods for object detection rely on a large amount of bounding box annotations. Collecting these annotations is laborious and costly, yet supervised models do not generalize well when testing on images from a different distribution. Domain adaptation provides a solution by adapting existing labels to the target testing data. However, a large gap between domains could make adaptation a challenging task, which leads to unstable training processes and sub-optimal results. In this paper, we propose to bridge the domain gap with an intermediate domain and progressively solve easier adaptation subtasks. This intermediate domain is constructed by translating the source images to mimic the ones in the target domain. To tackle the domain-shift problem, we adopt adversarial learning to align distributions at the feature level. In addition, a weighted task loss is applied to deal with unbalanced image quality in the intermediate domain. Experimental results show that our method performs favorably against the state-of-the-art method in terms of the performance on the target domain.

We address the problem of weakly supervised object localization where only image-level annotations are available for training object detectors. Numerous methods have been proposed to tackle this problem through mining object proposals. However, a substantial amount of noise in object proposals causes ambiguities for learning discriminative object models. Such approaches are sensitive to model initialization and often converge to undesirable local minimum solutions. In this paper, we propose to overcome these drawbacks by progressive representation adaptation with two main steps: 1) classification adaptation and 2) detection adaptation. In classification adaptation, we transfer a pre-trained network to a multi-label classification task for recognizing the presence of a certain object in an image. Through the classification adaptation step, the network learns discriminative representations that are specific to object categories of interest. In detection adaptation, we mine class-specific object proposals by exploiting two scoring strategies based on the adapted classification network. Class-specific proposal mining helps remove substantial noise from the background clutter and potential confusion from similar objects. We further refine these proposals using multiple instance learning and segmentation cues. Using these refined object bounding boxes, we fine-tune all the layer of the classification network and obtain a fully adapted detection network. We present detailed experimental validation on the PASCAL VOC and ILSVRC datasets. Experimental results demonstrate that our progressive representation adaptation algorithm performs favorably against the state-of-the-art methods.

Finding the informative subspaces of high-dimensional datasets is at the core of numerous applications in computer vision, where spectral-based subspace clustering is arguably the most widely studied method due to its strong empirical performance. Such algorithms first compute an affinity matrix to construct a self-representation for each sample using other samples as a dictionary. Sparsity and connectivity of the self-representation play important roles in effective subspace clustering. However, simultaneous optimization of both factors is difficult due to their conflicting nature, and most existing methods are designed to address only one factor. In this paper, we propose a post-processing technique to optimize both sparsity and connectivity by finding good neighbors. Good neighbors induce key connections among samples within a subspace and not only have large affinity coefficients but are also strongly connected to each other. We reassign the coefficients of the good neighbors and eliminate other entries to generate a new coefficient matrix. We show that the few good neighbors can effectively recover the subspace, and the proposed post-processing step of finding good neighbors is complementary to most existing subspace clustering algorithms. Experiments on five benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art methods with negligible additional computation cost.

We present an effective semi-supervised learning algorithm for single image dehazing. The proposed algorithm applies a deep Convolutional Neural Network (CNN) containing a supervised learning branch and an unsupervised learning branch. In the supervised branch, the deep neural network is constrained by the supervised loss functions, which are mean squared, perceptual, and adversarial losses. In the unsupervised branch, we exploit the properties of clean images via sparsity of dark channel and gradient priors to constrain the network. We train the proposed network on both the synthetic data and real-world images in an end-to-end manner. Our analysis shows that the proposed semi-supervised learning algorithm is not limited to synthetic training datasets and can be generalized well to real-world images. Extensive experimental results demonstrate that the proposed algorithm performs favorably against the state-of-the-art single image dehazing algorithms on both benchmark datasets and real-world images.

Dynamic scene blur is usually caused by object motion, depth variation as well as camera shake. Most existing methods usually solve this problem using image segmentation or fully end-to-end trainable deep convolutional neural networks by considering different object motions or camera shakes. However, these algorithms are less effective when there exist depth variations. In this work, we propose a deep neural convolutional network that exploits the depth map for dynamic scene deblurring. Given a blurred image, we first extract the depth map and adopt a depth refinement network to restore the edges and structure in the depth map. To effectively exploit the depth map, we adopt the spatial feature transform layer to extract depth features and fuse with the image features through scaling and shifting. Our image deblurring network thus learns to restore a clear image under the guidance of the depth map. With substantial experiments and analysis, we show that the depth information is crucial to the performance of the proposed model. Finally, extensive quantitative and qualitative evaluations demonstrate that the proposed model performs favorably against the state-of-the-art dynamic scene deblurring approaches as well as conventional depth-based deblurring algorithms.

Existing saliency models typically incorporate contexts holistically. However, for each pixel, usually only part of its context region contributes to saliency prediction, while other parts are likely either noise or distractions. In this paper, we propose a novel pixel-wise contextual attention network (PiCANet) to selectively attend to informative context locations at each pixel. The proposed PiCANet generates an attention map over the contextual region of each pixel and construct attentive contextual features via selectively incorporating the features of useful context locations. We present three formulations of the PiCANet via embedding the pixel-wise contextual attention mechanism into the pooling and convolution operations with attending to global or local contexts. All the three models are fully differentiable and can be integrated with convolutional neural networks with joint training. In this work, we introduce the proposed PiCANets into a U-Net model for salient object detection. The generated global and local attention maps can learn to incorporate global contrast and regional smoothness, which help localize and highlight salient objects more accurately and uniformly. Experimental results show that the proposed PiCANets perform effectively for saliency detection against the state-of-the-art methods. Furthermore, we demonstrate the effectiveness and generalization ability of the PiCANets on semantic segmentation and object detection with improved performance.

Referring expression is a special kind of verbal expression. The goal of referring expression is to refer to a particular object in some scenarios. Referring expression generation and comprehension are two inverse tasks within the field. Considering the critical role that visual attributes play in distinguishing the referred object from other objects, we propose an attribute-guided attention model to address the two tasks. In our proposed framework, attributes collected from referring expressions are used as explicit supervision signals on the generation and comprehension modules. The online predicted attributes of the visual object can benefit both tasks in two aspects: First, attributes can be directly embedded into the generation and comprehension modules, distinguishing the referred object as additional visual representations. Second, since attributes have their correspondence in both visual and textual space, an attribute-guided attention module is proposed as a bridging part to link the counterparts in visual representation and textual expression. Attention weights learned on both visual feature and word embeddings validate our motivation. We experiment on three standard datasets of RefCOCO, RefCOCO+ and RefCOCOg commonly used in this field. Both quantitative and qualitative results demonstrate the effectiveness of our proposed framework. The experimental results show significant improvements over baseline methods, and are favorably comparable to the state-of-the-art results. Further ablation study and analysis clearly demonstrate the contribution of each module, which could provide useful inspirations to the community.

Existing denoising methods typically restore clear results by aggregating pixels from the noisy input. Instead of relying on hand-crafted aggregation schemes, we propose to explicitly learn this process with deep neural networks. We present a spatial pixel aggregation network and learn the pixel sampling and averaging strategies for image denoising. The proposed model naturally adapts to image structures and can effectively improve the denoised results. Furthermore, we develop a spatio-temporal pixel aggregation network for video denoising to efficiently sample pixels across the spatio-temporal space. Our method is able to solve the misalignment issues caused by large motion in dynamic scenes. In addition, we introduce a new regularization term for effectively training the proposed video denoising model. We present extensive analysis of the proposed method and demonstrate that our model performs favorably against the state-of-the-art image and video denoising approaches on both synthetic and real-world data.

In recent years, convolutional neural networks (CNNs) have achieved great success in visual tracking. Most of existing methods train or fine-tune a binary classifier to distinguish the target from its background. However, they may suffer from the performance degradation due to insufficient training data. In this paper, we show that attribute information (e.g., illumination changes, occlusion and motion) in the context facilitates training an effective classifier for visual tracking. In particular, we design an attribute-based CNN with multiple branches, where each branch is responsible for classifying the target under a specific attribute. Such a design reduces the appearance diversity of the target under each attribute and thus requires less data to train the model. We combine all attributespecific features via ensemble layers to obtain more discriminative representations for the final target/background classification. The proposed method achieves favorable performance on the OTB100 dataset compared to state-of-the-art tracking methods. After being trained on the VOT datasets, the proposed network also shows a good generalization ability on the UAV-Traffic dataset, which has significantly different attributes and target appearances with the VOT datasets.

Video frame interpolation aims to synthesize nonexistent frames in-between the original frames. While significant advances have been made from the recent deep convolutional neural networks, the quality of interpolation is often reduced due to large object motion or occlusion. In this work, we propose a video frame interpolation method which explicitly detects the occlusion by exploring the depth information. Specifically, we develop a depth-aware flow projection layer to synthesize intermediate flows that preferably sample closer objects than farther ones. In addition, we learn hierarchical features to gather contextual information from neighboring pixels. The proposed model then warps the input frames, depth maps, and contextual features based on the optical flow and local interpolation kernels for synthesizing the output frame. Our model is compact, efficient, and fully differentiable. Quantitative and qualitative results demonstrate that the proposed model performs favorably against state-of-the-art frame interpolation methods on a wide variety of datasets. The source code and pre-trained model are available at https://github.com/baowenbo/DAIN.

Unsupervised domain adaptation algorithms aim to transfer the knowledge learned from one domain to another (e.g., synthetic to real images). The adapted representations often do not capture pixel-level domain shifts that are crucial for dense prediction tasks (e.g., semantic segmentation). In this paper, we present a novel pixel-wise adversarial domain adaptation algorithm. By leveraging image-to-image translation methods for data augmentation, our key insight is that while the translated images between domains may differ in styles, their predictions for the task should be consistent. We exploit this property and introduce a cross-domain consistency loss that enforces our adapted model to produce consistent predictions. Through extensive experimental results, we show that our method compares favorably against the state-of-the-art on a wide variety of unsupervised domain adaptation tasks.

Parts provide a good intermediate representation of objects that is robust with respect to camera, pose and appearance variations. Existing work on part segmentation is dominated by supervised approaches that rely on large amounts of manual annotations and also can not generalize to unseen object categories. We propose a self-supervised deep learning approach for part segmentation, where we devise several loss functions that aids in predicting part segments that are geometrically concentrated, robust to object variations and are also semantically consistent across different object instances. Extensive experiments on different types of image collections demonstrate that our approach can produce part segments that adhere to object boundaries and also more semantically consistent across object instances compared to existing self-supervised techniques.

In this paper, we introduce a new problem of manipulating a given video by inserting other videos into it. Our main task is, given an object video and a scene video, to insert the object video at a user-specified location in the scene video so that the resulting video looks realistic. We aim to handle different object motions and complex backgrounds without expensive segmentation annotations. As it is difficult to collect training pairs for this problem, we synthesize fake training pairs that can provide helpful supervisory signals when training a neural network with unpaired real data. The proposed network architecture can take both real and fake pairs as input and perform both supervised and unsupervised training in an adversarial learning scheme. To synthesize a realistic video, the network renders each frame based on the current input and previous frames. Within this framework, we observe that injecting noise into previous frames while generating the current frame stabilizes training. We conduct experiments on real-world videos in object tracking and person re-identification benchmark datasets. Experimental results demonstrate that the proposed algorithm is able to synthesize long sequences of realistic videos with a given object video inserted.

Existing deep trackers mainly use convolutional neural networks pre-trained for the generic object recognition task for representations. Despite demonstrated successes for numerous vision tasks, the contributions of using pre-trained deep features for visual tracking are not as significant as that for object recognition. The key issue is that in visual tracking the targets of interest can be arbitrary object class with arbitrary forms. As such, pre-trained deep features are less effective in modeling these targets of arbitrary forms for distinguishing them from the background. In this paper, we propose a novel scheme to learn target-aware features, which can better recognize the targets undergoing significant appearance variations than pre-trained deep features. To this end, we develop a regression loss and a ranking loss to guide the generation of target-active and scale-sensitive features. We identify the importance of each convolutional filter according to the back-propagated gradients and select the target-aware features based on activations for representing the targets. The target-aware features are integrated with a Siamese matching network for visual tracking. Extensive experimental results show that the proposed algorithm performs favorably against the state-of-the-art methods in terms of accuracy and speed.

We propose a high-quality photo-to-pencil translation method with fine-grained control over the drawing style. This is a challenging task due to multiple stroke types (e.g., outline and shading), structural complexity of pencil shading (e.g., hatching), and the lack of aligned training data pairs. To address these challenges, we develop a two-branch model that learns separate filters for generating sketchy outlines and tonal shading from a collection of pencil drawings. We create training data pairs by extracting clean outlines and tonal illustrations from original pencil drawings using image filtering techniques, and we manually label the drawing styles. In addition, our model creates different pencil styles (e.g., line sketchiness and shading style) in a user-controllable manner. Experimental results on different types of pencil drawings show that the proposed algorithm performs favorably against existing methods in terms of quality, diversity and user evaluations.

Given a random pair of images, a universal style transfer method extracts the feel from a reference image to synthesize an output based on the look of a content image. Recent algorithms based on second-order statistics, however, are either computationally expensive or prone to generate artifacts due to the trade-off between image quality and runtime performance. In this work, we present an approach for universal style transfer that learns the transformation matrix in a data-driven fashion. Our algorithm is efficient yet flexible to transfer different levels of styles with the same auto-encoder network. It also produces stable video style transfer results due to the preservation of the content affinity. In addition, we propose a linear propagation module to enable a feed-forward network for photo-realistic style transfer. We demonstrate the effectiveness of our approach on three tasks: artistic style, photo-realistic and video style transfer, with comparisons to state-of-the-art methods.

Affordance modeling plays an important role in visual understanding. In this paper, we aim to predict affordances of 3D indoor scenes, specifically what human poses are afforded by a given indoor environment, such as sitting on a chair or standing on the floor. In order to predict valid affordances and learn possible 3D human poses in indoor scenes, we need to understand the semantic and geometric structure of a scene as well as its potential interactions with a human. To learn such a model, a large-scale dataset of 3D indoor affordances is required. In this work, we build a fully automatic 3D pose synthesizer that fuses semantic knowledge from a large number of 2D poses extracted from TV shows as well as 3D geometric knowledge from voxel representations of indoor scenes. With the data created by the synthesizer, we introduce a 3D pose generative model to predict semantically plausible and physically feasible human poses within a given scene (provided as a single RGB, RGB-D, or depth image). We demonstrate that our human affordance prediction method consistently outperforms existing state-of-the-art methods.

Most conditional generation tasks expect diverse outputs given a single conditional context. However, conditional generative adversarial networks (cGANs) often focus on the prior conditional information and ignore the input noise vectors, which contribute to the output variations. Recent attempts to resolve the mode collapse issue for cGANs are usually task-specific and computationally expensive. In this work, we propose a simple yet effective regularization term to address the mode collapse issue for cGANs. The proposed method explicitly maximizes the ratio of the distance between generated images with respect to the corresponding latent codes, thus encouraging the generators to explore more minor modes during training. This mode seeking regularization term is readily applicable to various conditional generation tasks without imposing training overhead or modifying the original network structures. We validate the proposed algorithm on three conditional image synthesis tasks including categorical generation, image-to-image translation, and text-to-image synthesis with different baseline models. Both qualitative and quantitative results demonstrate the effectiveness of the proposed regularization method for improving diversity without loss of quality.

Joint filtering mainly uses an additional guidance image as a prior and transfers its structures to the target image in the filtering process. Different from existing algorithms that rely on locally linear models or hand-designed objective functions to extract the structural information from the guidance image, we propose a new joint filter based on a spatially variant linear representation model (SVLRM), where the target image is linearly represented by the guidance image. However, the SVLRM leads to a highly ill-posed problem. To estimate the linear representation coefficients, we develop an effective algorithm based on a deep convolutional neural network (CNN). The proposed deep CNN (constrained by the SVLRM) is able to estimate the spatially variant linear representation coefficients which are able to model the structural information of both the guidance and input images. We show that the proposed algorithm can be effectively applied to a variety of applications, including depth/RGB image upsampling and restoration, flash/no-flash image deblurring, natural image denoising, scale-aware filtering, etc. Extensive experimental results demonstrate that the proposed algorithm performs favorably against state-of-the-art methods that have been specially designed for each task.

A caricature is an artistic form of a person's picture in which certain striking characteristics are abstracted or exaggerated in order to create a humor or sarcasm effect. For numerous caricature related applications such as attribute recognition and caricature editing, face parsing is an essential pre-processing step that provides a complete facial structure understanding. However, current state-of-the-art face parsing methods require large amounts of labeled data on the pixel-level and such process for caricature is tedious and labor-intensive. For real photos, there are numerous labeled datasets for face parsing. Thus, we formulate caricature face parsing as a domain adaptation problem, where real photos play the role of the source domain, adapting to the target caricatures. Specifically, we first leverage a spatial transformer based network to enable shape domain shifts. A feed-forward style transfer network is then utilized to capture texture-level domain gaps. With these two steps, we synthesize face caricatures from real photos, and thus we can use parsing ground truths of the original photos to learn the parsing model. Experimental results on the synthetic and real caricatures demonstrate the effectiveness of the proposed domain adaptation algorithm. Code is available at: https://github.com/ZJULearning/CariFaceParsing.

Spatial audio is an essential medium to audiences for 3D visual and auditory experience. However, the recording devices and techniques are expensive or inaccessible to the general public. In this work, we propose a self-supervised audio spatialization network that can generate spatial audio given the corresponding video and monaural audio. To enhance spatialization performance, we use an auxiliary classifier to classify ground-truth videos and those with audio where the left and right channels are swapped. We collect a large-scale video dataset with spatial audio to validate the proposed method. Experimental results demonstrate the effectiveness of the proposed model on the audio spatialization task.

Dancing to music is an instinctive move by humans. Learning to model the music-to-dance generation process is, however, a challenging problem. It requires significant efforts to measure the correlation between music and dance as one needs to simultaneously consider multiple aspects, such as style and beat of both music and dance. Additionally, dance is inherently multimodal and various following movements of a pose at any moment are equally likely. In this paper, we propose a synthesis-by-analysis learning framework to generate dance from music. In the top-down analysis phase, we decompose a dance into a series of basic dance units, through which the model learns how to move. In the bottom-up synthesis phase, the model learns how to compose a dance by combining multiple basic dancing movements seamlessly according to input music. Experimental qualitative and quantitative results demonstrate that the proposed method can synthesize realistic, diverse, style-consistent, and beat-matching dances from music.

This paper proposes to learn reliable dense correspondence from videos in a self-supervised manner. Our learning process integrates two highly related tasks: tracking large image regions and establishing fine-grained pixel-level associations between consecutive video frames. We exploit the synergy between both tasks through a shared inter-frame affinity matrix, which simultaneously models transitions between video frames at both the region- and pixel-levels. While region-level localization helps reduce ambiguities in fine-grained matching by narrowing down search regions; fine-grained matching provides bottom-up features to facilitate region-level localization. Our method outperforms the state-of-the-art self-supervised methods on a variety of visual correspondence tasks, including video-object and part-segmentation propagation, keypoint tracking, and object tracking. Our self-supervised method even surpasses the fully-supervised affinity feature representation obtained from a ResNet-18 pre-trained on the ImageNet.

Video interpolation is an important problem in computer vision, which helps overcome the temporal limitation of camera sensors. Existing video interpolation methods usually assume uniform motion between consecutive frames and use linear models for interpolation, which cannot well approximate the complex motion in the real world. To address these issues, we propose a quadratic video interpolation method which exploits the acceleration information in videos. This method allows prediction with curvilinear trajectory and variable velocity, and generates more accurate interpolation results. For high-quality frame synthesis, we develop a flow reversal layer to estimate flow fields starting from the unknown target frame to the source frame. In addition, we present techniques for flow refinement. Extensive experiments demonstrate that our approach performs favorably against the existing linear models on a wide variety of video datasets.

Deep Convolutional Neural Networks (CNNs) have been repeatedly proven to perform well on image classification tasks. Object detection methods, however, are still in need of significant improvements. In this paper, we propose a new framework called Ventral-Dorsal Networks (VDNets) which is inspired by the structure of the human visual system. Roughly, the visual input signal is analyzed along two separate neural streams, one in the temporal lobe and the other in the parietal lobe. The coarse functional distinction between these streams is between object recognition — the "what" of the signal - and extracting location related information — the "where" of the signal. The ventral pathway from primary visual cortex, entering the temporal lobe, is dominated by "what" information, while the dorsal pathway, into the parietal lobe, is dominated by "where" information. Inspired by this structure, we propose the integration of a "Ventral Network" and a "Dorsal Network", which are complementary. Information about object identity can guide localization, and location information can guide attention to relevant image regions, improving object recognition. This new dual network framework sharpens the focus of object detection. Our experimental results reveal that the proposed method outperforms state-of-the-art object detection approaches on PASCAL VOC 2007 by 8% (mAP) and PASCAL VOC 2012 by 3% (mAP). Moreover, a comparison of techniques on Yearbook images displays substantial qualitative and quantitative benefits of VDNet.

To date, top-performing optical flow estimation methods only take pairs of consecutive frames into account. While elegant and appealing, the idea of using more than two frames has not yet produced state-of-the-art results. We present a simple, yet effective fusion approach for multi-frame optical flow that benefits from longer-term temporal cues. Our method first warps the optical flow from previous frames to the current, thereby yielding multiple plausible estimates. It then fuses the complementary information carried by these estimates into a new optical flow field. At the time of writing, our method ranks first among published results in the MPI Sintel and KITTI 2015 benchmarks. Our models will be available on https://github.com/NVlabs/PWC-Net.

Existing visual tracking methods usually localize a target object with a bounding box, in which the performance of the foreground object trackers or detectors is often affected by the inclusion of background clutter. To handle this problem, we learn a patch-based graph representation for visual tracking. The tracked object is modeled by with a graph by taking a set of non-overlapping image patches as nodes, in which the weight of each node indicates how likely it belongs to the foreground and edges are weighted for indicating the appearance compatibility of two neighboring nodes. This graph is dynamically learned and applied in object tracking and model updating. During the tracking process, the proposed algorithm performs three main steps in each frame. First, the graph is initialized by assigning binary weights of some image patches to indicate the object and background patches according to the predicted bounding box. Second, the graph is optimized to refine the patch weights by using a novel alternating direction method of multipliers. Third, the object feature representation is updated by imposing the weights of patches on the extracted image features. The object location is predicted by maximizing the classification score in the structured support vector machine. Extensive experiments show that the proposed tracking algorithm performs well against the state-of-the-art methods on large-scale benchmark datasets.

Visual tracking is challenging as target objects often undergo significant appearance changes caused by deformation, abrupt motion, background clutter and occlusion. In this paper, we propose to exploit the rich hierarchical features of deep convolutional neural networks to improve the accuracy and robustness of visual tracking. Deep neural networks trained on object recognition datasets consist of multiple convolutional layers. These layers encode target appearance with different levels of abstraction. For example, the outputs of the last convolutional layers encode the semantic information of targets and such representations are invariant to significant appearance variations. However, their spatial resolutions are too coarse to precisely localize the target. In contrast, features from earlier convolutional layers provide more precise localization but are less invariant to appearance changes. We interpret the hierarchical features of convolutional layers as a nonlinear counterpart of an image pyramid representation and explicitly exploit these multiple levels of abstraction to represent target objects. Specifically, we learn adaptive correlation filters on the outputs from each convolutional layer to encode the target appearance. We infer the maximum response of each layer to locate targets in a coarse-to-fine manner. To further handle the issues with scale estimation and re-detecting target objects from tracking failures caused by heavy occlusion or out-of-the-view movement, we conservatively learn another correlation filter, that maintains a long-term memory of target appearance, as a discriminative classifier. We apply the classifier to two types of object proposals: (1) proposals with a small step size and tightly around the estimated location for scale estimation; and (2) proposals with large step size and across the whole image for target re-detection. Extensive experimental results on large-scale benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art tracking methods.

Convolutional neural networks have recently demonstrated high-quality reconstruction for single image super-resolution. However, existing methods often require a large number of network parameters and entail heavy computational loads at runtime for generating high-accuracy super-resolution results. In this paper, we propose the deep Laplacian Pyramid Super-Resolution Network for fast and accurate image super-resolution. The proposed network progressively reconstructs the sub-band residuals of high-resolution images at multiple pyramid levels. In contrast to existing methods that involve the bicubic interpolation for pre-processing (which results in large feature maps), the proposed method directly extracts features from the low-resolution input space and thereby entails low computational loads. We train the proposed network with deep supervision using the robust Charbonnier loss functions and achieve high-quality image reconstruction. Furthermore, we utilize the recursive layers to share parameters across as well as within pyramid levels, and thus drastically reduce the number of parameters. Extensive quantitative and qualitative evaluations on benchmark datasets show that the proposed algorithm performs favorably against the state-of-the-art methods in terms of run-time and image quality.

In this paper, we consider the problem of estimating the head pose and body orientation of a person from a low-resolution image. Under this setting, it is difficult to reliably extract facial features or detect body parts. We propose a convolutional random projection forest (CRPforest) algorithm for these tasks. A convolutional random projection network (CRPnet) is used at each node of the forest. It maps an input image to a high-dimensional feature space using a rich filter bank. The filter bank is designed to generate sparse responses so that they can be efficiently computed by compressive sensing. A sparse random projection matrix can capture most essential information contained in the filter bank without using all the filters in it. Therefore, the CRPnet is fast, e.g., it requires to process an image of pixels, due to the small number of convolutions (e.g., 0.01 percent of a layer of a neural network) at the expense of less than 2 percent accuracy. The overall forest estimates head and body pose well on benchmark datasets, e.g., over 98 percent on the HIIT dataset, while requiring without using a GPU. Extensive experiments on challenging datasets show that the proposed algorithm performs favorably against the state-of-the-art methods in low-resolution images with noise, occlusion, and motion blur.

Joint image filters leverage the guidance image as a prior and transfer the structural details from the guidance image to the target image for suppressing noise or enhancing spatial resolution. Existing methods either rely on various explicit filter constructions or hand-designed objective functions, thereby making it difficult to understand, improve, and accelerate these filters in a coherent framework. In this paper, we propose a learning-based approach for constructing joint filters based on Convolutional Neural Networks. In contrast to existing methods that consider only the guidance image, the proposed algorithm can selectively transfer salient structures that are consistent with both guidance and target images. We show that the model trained on a certain type of data, e.g., RGB and depth images, generalizes well to other modalities, e.g., flash/non-Flash and RGB/NIR images. We validate the effectiveness of the proposed joint filter through extensive experimental evaluations with state-of-the-art methods.

In the fields of computer vision and graphics, keypoint-based object tracking is a fundamental and challenging problem, which is typically formulated in a spatio-temporal context modeling framework. However, many existing keypoint trackers are incapable of effectively modeling and balancing the following three aspects in a simultaneous manner: temporal model coherence across frames, spatial model consistency within frames, and discriminative feature construction. To address this problem, we propose a robust keypoint tracker based on spatio-temporal multi-task structured output optimization driven by discriminative metric learning. Consequently, temporal model coherence is characterized by multi-task structured keypoint model learning over several adjacent frames; spatial model consistency is modeled by solving a geometric verification based structured learning problem; discriminative feature construction is enabled by metric learning to ensure the intra-class compactness and inter-class separability. To achieve the goal of effective object tracking, we jointly optimize the above three modules in a spatio-temporal multi-task learning scheme. Furthermore, we incorporate this joint learning scheme into both single-object and multi-object tracking scenarios, resulting in robust tracking results. Experiments over several challenging datasets have justified the effectiveness of our single-object and multi-object trackers against the state-of-the-art.

Human faces are one interesting object class with numerous applications. While significant progress has been made in the generic deblurring problem, existing methods are less effective for blurry face images. The success of the state-of-the-art image deblurring algorithms stems mainly from implicit or explicit restoration of salient edges for kernel estimation. However, existing methods are less effective as only few edges can be restored from blurry face images for kernel estimation. In this paper, we address the problem of deblurring face images by exploiting facial structures. We propose a deblurring algorithm based on an exemplar dataset without using coarse-to-fine strategies or heuristic edge selections. In addition, we develop a convolutional neural network to restore sharp edges from blurry images for deblurring. Extensive experiments against the state-of-the-art methods demonstrate the effectiveness of the proposed algorithm for deblurring face images. In addition, we show that the proposed algorithms can be applied to image deblurring for other object classes.

Convolutional Neural Networks (CNNs) have been applied to visual tracking with demonstrated success in recent years. Most CNN-based trackers utilize hierarchical features extracted from a certain layer to represent the target. However, features from a certain layer are not always effective for distinguishing the target object from the backgrounds especially in the presence of complicated interfering factors (e.g., heavy occlusion, background clutter, illumination variation, and shape deformation). In this work, we propose a CNN-based tracking algorithm which hedges deep features from different CNN layers to better distinguish target objects and background clutters. Correlation filters are applied to feature maps of each CNN layer to construct a weak tracker, and all weak trackers are hedged into a strong one. For robust visual tracking, we propose a hedge method to adaptively determine weights of weak classifiers by considering both the difference between the historical as well as instantaneous performance, and the difference among all weak trackers over time. In addition, we design a Siamese network to define the loss of each weak tracker for the proposed hedge method. Extensive experiments on large benchmark datasets demonstrate the effectiveness of the proposed algorithm against the state-of-the-art tracking methods.

In this paper, we propose a multi-task correlation particle filter (MCPF) for robust visual tracking. We first present the multi-task correlation filter (MCF) that takes the interdependencies among different object parts and features into account to learn the correlation filters jointly. Next, the proposed MCPF is introduced to exploit and complement the strength of a MCF and a particle filter. Compared with existing tracking methods based on correlation filters and particle filters, the proposed MCPF enjoys several merits. First, it exploits the interdependencies among different features to derive the correlation filters jointly, and makes the learned filters complement and enhance each other to obtain consistent responses. Second, it handles partial occlusion via a part-based representation, and exploits the intrinsic relationship among local parts via spatial constraints to preserve object structure and learn the correlation filters jointly. Third, it effectively handles large scale variation via a sampling scheme by drawing particles at different scales for target object state estimation. Fourth, it shepherds the sampled particles toward the modes of the target state distribution via the MCF, and effectively covers object states well using fewer particles than conventional particle filters, thereby resulting in robust tracking performance and low computational cost. Extensive experimental results on four challenging benchmark datasets demonstrate that the proposed MCPF tracking algorithm performs favorably against the state-of-the-art methods.

Sparse representations have been applied to visual tracking by finding the best candidate region with minimal reconstruction error based on a set of target templates. However, most existing sparse trackers only consider holistic or local representations and do not make full use of the intrinsic structure among and inside target candidate regions, thereby making them less effective when similar objects appear at close proximity or under occlusion. In this paper, we propose a novel structural sparse representation, which not only exploits the intrinsic relationships among target candidate regions and local patches to learn their representations jointly, but also preserves the spatial structure among the local patches inside each target candidate region. For robust visual tracking, we take outliers resulting from occlusion and noise into account when searching for the best target region. Constructed within a Bayesian filtering framework, we show that the proposed algorithm accommodates most existing sparse trackers with respective merits. The formulated problem can be efficiently solved using an accelerated proximal gradient method that yields a sequence of closed form updates. Qualitative and quantitative evaluations on challenging benchmark datasets demonstrate that the proposed tracking algorithm performs favorably against several state-of-the-art methods.

For visual tracking methods based on kernel support vector machines (SVMs), data sampling is usually adopted to reduce the computational cost in training. In addition, budgeting of support vectors is required for computational efficiency. Instead of sampling and budgeting, recently the circulant matrix formed by dense sampling of translated image patches has been utilized in kernel correlation filters for fast tracking. In this paper, we derive an equivalent formulation of a SVM model with the circulant matrix expression and present an efficient alternating optimization method for visual tracking. We incorporate the discrete Fourier transform with the proposed alternating optimization process, and pose the tracking problem as an iterative learning of support correlation filters (SCFs). In the fully-supervision setting, our SCF can find the globally optimal solution with real-time performance. For a given circulant data matrix with$n^2$samples of$n \times n$pixels, the computational complexity of the proposed algorithm is$O(n^2\; \log n)$whereas that of the standard SVM-based approaches is at least$O(n^4)$. In addition, we extend the SCF-based tracking algorithm with multi-channel features, kernel functions, and scale-adaptive approaches to further improve the tracking performance. Experimental results on a large benchmark dataset show that the proposed SCF-based algorithms perform favorably against the state-of-the-art tracking methods in terms of accuracy and speed.

Object proposal algorithms have been demonstrated to be very successful in accelerating object detection process. High object localization quality and detection recall can be obtained using thousands of proposals. However, the performance with a small number of proposals is still unsatisfactory. This paper demonstrates that the performance of a few proposals can be significantly improved with the minimal human interaction-a single touch point. To this end, we first generate hierarchical superpixels using an efficient tree-organized structure as our initial object proposals, and then select only a few proposals from them by learning an effective Convolutional neural network for objectness ranking. We explore and design an architecture to integrate human interaction with the global information of the whole image for objectness scoring, which is able to significantly improve the performance with a minimum number of object proposals. Extensive experiments show the proposed method outperforms all the state-of-the-art methods for locating the meaningful object with the touch point constraint. Furthermore, the proposed method is extended for video. By combining with the novel interactive motion segmentation cue for generating hierarchical superpixels, the performance on a single proposal is satisfactory and can be used in the interactive vision systems, such as selecting the input of a real-time tracking system.

Camera sensors often fail to capture clear images or videos in a poorly lit environment. In this paper, we propose a trainable hybrid network to enhance the visibility of such degraded images. The proposed network consists of two distinct streams to simultaneously learn the global content and the salient structures of the clear image in a unified network. More specifically, the content stream estimates the global content of the low-light input through an encoder-decoder network. However, the encoder in the content stream tends to lose some structure details. To remedy this, we propose a novel spatially variant recurrent neural network (RNN) as an edge stream to model edge details, with the guidance of another auto-encoder. The experimental results show that the proposed network favorably performs against the state-of-the-art low-light image enhancement algorithms.

The recent years have witnessed advances in parallel algorithms for large-scale optimization problems. Notwithstanding the demonstrated success, existing algorithms that parallelize over features are usually limited by divergence issues under high parallelism or require data preprocessing to alleviate these problems. In this paper, we propose a Parallel Coordinate Descent algorithm using approximate Newton steps (PCDN) that is guaranteed to converge globally without data preprocessing. The key component of the PCDN algorithm is the high-dimensional line search, which guarantees the global convergence with high parallelism. The PCDN algorithm randomly partitions the feature set into b subsets/bundles of size P, and sequentially processes each bundle by first computing the descent directions for each feature in parallel and then conducting P-dimensional line search to compute the step size. We show that: 1) the PCDN algorithm is guaranteed to converge globally despite increasing parallelism and 2) the PCDN algorithm converges to the specified accuracy ϵ within the limited iteration number of Tϵ, and Tϵdecreases with increasing parallelism. In addition, the data transfer and synchronization cost of the P-dimensional line search can be minimized by maintaining intermediate quantities. For concreteness, the proposed PCDN algorithm is applied to L1-regularized logistic regression and L1-regularized L2-loss support vector machine problems. Experimental evaluations on seven benchmark data sets show that the PCDN algorithm exploits parallelism well and outperforms the state-of-the-art methods.

Person reidentification is of great importance in visual surveillance and multiperson tracking across multiple camera views. Two fundamental problems are critical for person reidentification: 1) how to account for appearance variation or feature transformation caused by viewpoint changes and 2) how to learn a discriminative distance metric for reidentification. In this paper, we propose an algorithm in which both feature transformation and metric learning are exploited and jointly optimized. We learn local models from subsets of training samples with regularization imposed by the global model which is trained among the entire data set. The learned local models enhance the discriminative strength and generalization ability. Experimental results on the Viewpoint Invariant PEdestrian Eecognition, Queen Mary University of London ground reidentification, CUHK01, and CUHK03 benchmark data sets show that the proposed sample-specific view-invariant approach performs favorably against the state-of-the-art person reidentification methods.

We study domain-specific video streaming. Specifically, we target a streaming setting where the videos to be streamed from a server to a client are all in the same domain and they have to be compressed to a small size for low-latency transmission. Several popular video streaming services, such as the video game streaming services of GeForce Now and Twitch, fall in this category. While conventional video compression standards such as H.264 are commonly used for this task, we hypothesize that one can leverage the property that the videos are all in the same domain to achieve better video quality. Based on this hypothesis, we propose a novel video compression pipeline. Specifically, we first apply H.264 to compress domain-specific videos. We then train a novel binary autoencoder to encode the leftover domain-specific residual information frame-by-frame into binary representations. These binary representations are then compressed and sent to the client together with the H.264 stream. In our experiments, we show that our pipeline yields consistent gains over standard H.264 compression across several benchmark datasets while using the same channel bandwidth.

Affective image understanding has been extensively studied in the last decade since more and more users express emotion via visual contents. While current algorithms based on convolutional neural networks aim to distinguish emotional categories in a discrete label space, the task is inherently ambiguous. This is mainly because emotional labels with the same polarity (i.e., positive or negative) are highly related, which is different from concrete object concepts such as cat, dog and bird. To the best of our knowledge, few methods focus on leveraging such characteristic of emotions for affective image understanding. In this work, we address the problem of understanding affective images via deep metric learning and propose a multi-task deep framework to optimize both retrieval and classification goals. We propose the sentiment constraints adapted from the triplet constraints, which are able to explore the hierarchical relation of emotion labels. We further exploit the sentiment vector as an effective representation to distinguish affective images utilizing the texture representation derived from convolutional layers. Extensive evaluations on four widely-used affective datasets, i.e., Flickr and Instagram, IAPSa, Art Photo, and Abstract Paintings, demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both affective image retrieval and classification tasks.

Online video object segmentation is a challenging task as it entails to process the image sequence timely and accurately. To segment a target object through the video, numerous CNN-based methods have been developed by heavily finetuning on the object mask in the first frame, which is time-consuming for online applications. In this paper, we propose a fast and accurate video object segmentation algorithm that can immediately start the segmentation process once receiving the images. We first utilize a part-based tracking method to deal with challenging factors such as large deformation, occlusion, and cluttered background. Based on the tracked bounding boxes of parts, we construct a region-of-interest segmentation network to generate part masks. Finally, a similarity-based scoring function is adopted to refine these object parts by comparing them to the visual information in the first frame. Our method performs favorably against state-of-the-art algorithms in accuracy on the DAVIS benchmark dataset, while achieving much faster runtime performance.

Given two consecutive frames, video interpolation aims at generating intermediate frame(s) to form both spatially and temporally coherent video sequences. While most existing methods focus on single-frame interpolation, we propose an end-to-end convolutional neural network for variable-length multi-frame video interpolation, where the motion interpretation and occlusion reasoning are jointly modeled. We start by computing bi-directional optical flow between the input images using a U-Net architecture. These flows are then linearly combined at each time step to approximate the intermediate bi-directional optical flows. These approximate flows, however, only work well in locally smooth regions and produce artifacts around motion boundaries. To address this shortcoming, we employ another U-Net to refine the approximated flow and also predict soft visibility maps. Finally, the two input images are warped and linearly fused to form each intermediate frame. By applying the visibility maps to the warped images before fusion, we exclude the contribution of occluded pixels to the interpolated intermediate frame to avoid artifacts. Since none of our learned network parameters are time-dependent, our approach is able to produce as many intermediate frames as needed. To train our network, we use 1,132 240-fps video clips, containing 300K individual video frames. Experimental results on several datasets, predicting different numbers of interpolated frames, demonstrate that our approach performs consistently better than existing methods.

We present an effective blind image deblurring method based on a data-driven discriminative prior. Our work is motivated by the fact that a good image prior should favor clear images over blurred ones. In this work, we formulate the image prior as a binary classifier which can be achieved by a deep convolutional neural network (CNN). The learned prior is able to distinguish whether an input image is clear or not. Embedded into the maximum a posterior (MAP) framework, it helps blind deblurring in various scenarios, including natural, face, text, and low-illumination images. However, it is difficult to optimize the deblurring method with the learned image prior as it involves a non-linear CNN. Therefore, we develop an efficient numerical approach based on the half-quadratic splitting method and gradient decent algorithm to solve the proposed model. Furthermore, the proposed model can be easily extended to non-uniform deblurring. Both qualitative and quantitative experimental results show that our method performs favorably against state-of-the-art algorithms as well as domain-specific image deblurring approaches.

Discriminative Correlation Filters (DCF) are efficient in visual tracking but suffer from unwanted boundary effects. Spatially Regularized DCF (SRDCF) has been suggested to resolve this issue by enforcing spatial penalty on DCF coefficients, which, inevitably, improves the tracking performance at the price of increasing complexity. To tackle online updating, SRDCF formulates its model on multiple training images, further adding difficulties in improving efficiency. In this work, by introducing temporal regularization to SRDCF with single sample, we present our spatialtemporal regularized correlation filters (STRCF). The STRCF formulation can not only serve as a reasonable approximation to SRDCF with multiple training samples, but also provide a more robust appearance model than SRDCF in the case of large appearance variations. Besides, it can be efficiently solved via the alternating direction method of multipliers (ADMM). By incorporating both temporal and spatial regularization, our STRCF can handle boundary effects without much loss in efficiency and achieve superior performance over SRDCF in terms of accuracy and speed. Compared with SRDCF, STRCF with hand-crafted features provides a 5× speedup and achieves a gain of 5.4% and 3.6% AUC score on OTB-2015 and Temple-Color, respectively. Moreover, STRCF with deep features also performs favorably against state-of-the-art trackers and achieves an AUC score of 68.3% on OTB-2015.

Contexts play an important role in the saliency detection task. However, given a context region, not all contextual information is helpful for the final task. In this paper, we propose a novel pixel-wise contextual attention network, i.e., the PiCANet, to learn to selectively attend to informative context locations for each pixel. Specifically, for each pixel, it can generate an attention map in which each attention weight corresponds to the contextual relevance at each context location. An attended contextual feature can then be constructed by selectively aggregating the contextual information. We formulate the proposed PiCANet in both global and local forms to attend to global and local contexts, respectively. Both models are fully differentiable and can be embedded into CNNs for joint training. We also incorporate the proposed models with the U-Net architecture to detect salient objects. Extensive experiments show that the proposed PiCANets can consistently improve saliency detection performance. The global and local PiCANets facilitate learning global contrast and homogeneousness, respectively. As a result, our saliency model can detect salient objects more accurately and uniformly, thus performing favorably against the state-of-the-art methods.

In this paper, we propose a general dual convolutional neural network (DualCNN) for low-level vision problems, e.g., super-resolution, edge-preserving filtering, deraining and dehazing. These problems usually involve the estimation of two components of the target signals: structures and details. Motivated by this, our proposed DualCNN consists of two parallel branches, which respectively recovers the structures and details in an end-to-end manner. The recovered structures and details can generate the target signals according to the formation model for each particular application. The DualCNN is a flexible framework for low-level vision tasks and can be easily incorporated into existing CNNs. Experimental results show that the DualCNN can be effectively applied to numerous low-level vision tasks with favorable performance against the state-of-the-art methods.

In this paper, we propose an efficient algorithm to directly restore a clear image from a hazy input. The proposed algorithm hinges on an end-to-end trainable neural network that consists of an encoder and a decoder. The encoder is exploited to capture the context of the derived input images, while the decoder is employed to estimate the contribution of each input to the final dehazed result using the learned representations attributed to the encoder. The constructed network adopts a novel fusion-based strategy which derives three inputs from an original hazy image by applying White Balance (WB), Contrast Enhancing (CE), and Gamma Correction (GC). We compute pixel-wise confidence maps based on the appearance differences between these different inputs to blend the information of the derived inputs and preserve the regions with pleasant visibility. The final dehazed image is yielded by gating the important features of the derived inputs. To train the network, we introduce a multi-scale approach such that the halo artifacts can be avoided. Extensive experimental results on both synthetic and real-world images demonstrate that the proposed algorithm performs favorably against the state-of-the-art algorithms.

Visual events are usually accompanied by sounds in our daily lives. We pose the question: Can the machine learn the correspondence between visual scene and the sound, and localize the sound source only by observing sound and visual scene pairs like human? In this paper, we propose a novel unsupervised algorithm to address the problem of localizing the sound source in visual scenes. A two-stream network structure which handles each modality, with attention mechanism is developed for sound source localization. Moreover, although our network is formulated within the unsupervised learning framework, it can be extended to a unified architecture with a simple modification for the supervised and semi-supervised learning settings as well. Meanwhile, a new sound source dataset is developed for performance evaluation. Our empirical evaluation shows that the unsupervised method eventually go through false conclusion in some cases. We also show that even with a few supervision, i.e., semi-supervised setup, false conclusion is able to be corrected effectively.

In this paper, we present an effective and efficient face deblurring algorithm by exploiting semantic cues via deep convolutional neural networks (CNNs). As face images are highly structured and share several key semantic components (e.g., eyes and mouths), the semantic information of a face provides a strong prior for restoration. As such, we propose to incorporate global semantic priors as input and impose local structure losses to regularize the output within a multi-scale deep CNN. We train the network with perceptual and adversarial losses to generate photo-realistic results and develop an incremental training strategy to handle random blur kernels in the wild. Quantitative and qualitative evaluations demonstrate that the proposed face deblurring algorithm restores sharp images with more facial details and performs favorably against state-of-the-art methods in terms of restoration quality, face recognition and execution speed.

The tracking-by-detection framework consists of two stages, i.e., drawing samples around the target object in the first stage and classifying each sample as the target object or as background in the second stage. The performance of existing trackers using deep classification networks is limited by two aspects. First, the positive samples in each frame are highly spatially overlapped, and they fail to capture rich appearance variations. Second, there exists extreme class imbalance between positive and negative samples. This paper presents the VITAL algorithm to address these two problems via adversarial learning. To augment positive samples, we use a generative network to randomly generate masks, which are applied to adaptively dropout input features to capture a variety of appearance changes. With the use of adversarial learning, our network identifies the mask that maintains the most robust features of the target objects over a long temporal span. In addition, to handle the issue of class imbalance, we propose a high-order cost sensitive loss to decrease the effect of easy negative samples to facilitate training the classification network. Extensive experiments on benchmark datasets demonstrate that the proposed tracker performs favorably against state-of-the-art approaches.

We present a network architecture for processing point clouds that directly operates on a collection of points represented as a sparse set of samples in a high-dimensional lattice. Naively applying convolutions on this lattice scales poorly, both in terms of memory and computational cost, as the size of the lattice increases. Instead, our network uses sparse bilateral convolutional layers as building blocks. These layers maintain efficiency by using indexing structures to apply convolutions only on occupied parts of the lattice, and allow flexible specifications of the lattice structure enabling hierarchical and spatially-aware feature learning, as well as joint 2D-3D reasoning. Both point-based and image-based representations can be easily incorporated in a network with such layers and the resulting model can be trained in an end-to-end manner. We present results on 3D segmentation tasks where our approach outperforms existing state-of-the-art techniques.

For visual tracking, an ideal filter learned by the correlation filter (CF) method should take both discrimination and reliability information. However, existing attempts usually focus on the former one while pay less attention to reliability learning. This may make the learned filter be dominated by the unexpected salient regions on the feature map, thereby resulting in model degradation. To address this issue, we propose a novel CF-based optimization problem to jointly model the discrimination and reliability information. First, we treat the filter as the element-wise product of a base filter and a reliability term. The base filter is aimed to learn the discrimination information between the target and backgrounds, and the reliability term encourages the final filter to focus on more reliable regions. Second, we introduce a local response consistency regular term to emphasize equal contributions of different regions and avoid the tracker being dominated by unreliable regions. The proposed optimization problem can be solved using the alternating direction method and speeded up in the Fourier domain. We conduct extensive experiments on the OTB-2013, OTB-2015 and VOT-2016 datasets to evaluate the proposed tracker. Experimental results show that our tracker performs favorably against other state-of-the-art trackers.

In this paper, we analyze the spatial information of deep features, and propose two complementary regressions for robust visual tracking. First, we propose a kernelized ridge regression model wherein the kernel value is defined as the weighted sum of similarity scores of all pairs of patches between two samples. We show that this model can be formulated as a neural network and thus can be efficiently solved. Second, we propose a fully convolutional neural network with spatially regularized kernels, through which the filter kernel corresponding to each output channel is forced to focus on a specific region of the target. Distance transform pooling is further exploited to determine the effectiveness of each output channel of the convolution layer. The outputs from the kernelized ridge regression model and the fully convolutional neural network are combined to obtain the ultimate response. Experimental results on two benchmark datasets validate the effectiveness of the proposed method.

Convolutional neural network-based approaches for semantic segmentation rely on supervision with pixel-level ground truth, but may not generalize well to unseen image domains. As the labeling process is tedious and labor intensive, developing algorithms that can adapt source ground truth labels to the target domain is of great interest. In this paper, we propose an adversarial learning method for domain adaptation in the context of semantic segmentation. Considering semantic segmentations as structured outputs that contain spatial similarities between the source and target domains, we adopt adversarial learning in the output space. To further enhance the adapted model, we construct a multi-level adversarial network to effectively perform output space domain adaptation at different feature levels. Extensive experiments and ablation study are conducted under various domain adaptation settings, including synthetic-to-real and cross-city scenarios. We show that the proposed method performs favorably against the state-of-the-art methods in terms of accuracy and visual quality.

Superpixel segmentation has been widely used in many computer vision tasks. Existing superpixel algorithms are mainly based on hand-crafted features, which often fail to preserve weak object boundaries. In this work, we leverage deep neural networks to facilitate extracting superpixels from images. We show a simple integration of deep features with existing superpixel algorithms does not result in better performance as these features do not model segmentation. Instead, we propose a segmentation-aware affinity learning approach for superpixel segmentation. Specifically, we propose a new loss function that takes the segmentation error into account for affinity learning. We also develop the Pixel Affinity Net for affinity prediction. Extensive experimental results show that the proposed algorithm based on the learned segmentation-aware loss performs favorably against the state-of-the-art methods. We also demonstrate the use of the learned superpixels in numerous vision applications with consistent improvements.

Automatic assessment of sentiment from visual content has gained considerable attention with the increasing tendency of expressing opinions on-line. In this paper, we solve the problem of visual sentiment analysis using the high-level abstraction in the recognition process. Existing methods based on convolutional neural networks learn sentiment representations from the holistic image appearance. However, different image regions can have a different influence on the intended expression. This paper presents a weakly supervised coupled convolutional network with two branches to leverage the localized information. The first branch detects a sentiment specific soft map by training a fully convolutional network with the cross spatial pooling strategy, which only requires image-level labels, thereby significantly reducing the annotation burden. The second branch utilizes both the holistic and localized information by coupling the sentiment map with deep features for robust classification. We integrate the sentiment detection and classification branches into a unified deep framework and optimize the network in an end-to-end manner. Extensive experiments on six benchmark datasets demonstrate that the proposed method performs favorably against the state-of-the-art methods for visual sentiment analysis.

Due to the spatially variant blur caused by camera shake and object motions under different scene depths, deblurring images captured from dynamic scenes is challenging. Although recent works based on deep neural networks have shown great progress on this problem, their models are usually large and computationally expensive. In this paper, we propose a novel spatially variant neural network to address the problem. The proposed network is composed of three deep convolutional neural networks (CNNs) and a recurrent neural network (RNN). RNN is used as a deconvolution operator performed on feature maps extracted from the input image by one of the CNNs. Another CNN is used to learn the weights for the RNN at every location. As a result, the RNN is spatially variant and could implicitly model the deblurring process with spatially variant kernels. The third CNN is used to reconstruct the final deblurred feature maps into restored image. The whole network is end-to-end trainable. Our analysis shows that the proposed network has a large receptive field even with a small model size. Quantitative and qualitative evaluations on public datasets demonstrate that the proposed method performs favorably against state-of-the-art algorithms in terms of accuracy, speed, and model size.

The 1st Workshop and Challenge on Comprehensive Video Understanding in the Wild, dubbed CoVieW'18, is held in Seoul, Korea on October 22, 2018, in conjuction with ACM Multimedia 2018. The workshop aims to solve the joint and comprehensive understanding problem in untrimmed videos with a particular emphasis on joint action and scene recognition. The workshop encourages researchers to participate in joint action and scene recognition challenge in untrimmed videos and to report their results. The workshop program includes 1 keynote speech, 2 invited speakers, 6 regular and challenge papers. The developments made in the workshop will deliver a step change in a variety of video applications.

Learning to insert an object instance into an image in a semantically coherent manner is a challenging and interesting problem. Solving it requires (a) determining a location to place an object in the scene and (b) determining its appearance at the location. Such an object insertion model can potentially facilitate numerous image editing and scene parsing applications. In this paper, we propose an end-to-end trainable neural network for the task of inserting an object instance mask of a specified class into the semantic label map of an image. Our network consists of two generative modules where one determines where the inserted object mask should be (i.e., location and scale) and the other determines what the object mask shape (and pose) should look like. The two modules are connected together via a spatial transformation network and jointly trained. We devise a learning procedure that leverage both supervised and unsupervised data and show our model can insert an object at diverse locations with various appearances. We conduct extensive experimental validations with comparisons to strong baselines to verify the effectiveness of the proposed network.

Visual attention, derived from cognitive neuroscience, facilitates human perception on the most pertinent subset of the sensory data. Recently, significant efforts have been made to exploit attention schemes to advance computer vision systems. For visual tracking, it is often challenging to track target objects undergoing large appearance changes. Attention maps facilitate visual tracking by selectively paying attention to temporal robust features. Existing tracking-by-detection approaches mainly use additional attention modules to generate feature weights as the classifiers are not equipped with such mechanisms. In this paper, we propose a reciprocative learning algorithm to exploit visual attention for training deep classifiers. The proposed algorithm consists of feed-forward and backward operations to generate attention maps, which serve as regularization terms coupled with the original classification loss function for training. The deep classifier learns to attend to the regions of target objects robust to appearance changes. Extensive experiments on large-scale benchmark datasets show that the proposed attentive tracking method performs favorably against the state-of-the-art approaches.

In this paper, we present a deep convolutional neural network to capture the inherent properties of image degradation, which can handle different kernels and saturated pixels in a unified framework. The proposed neural network is motivated by the low-rank property of pseudo-inverse kernels. We first compute a generalized low-rank approximation for a large number of blur kernels, and then use separable filters to initialize the convolutional parameters in the network. Our analysis shows that the estimated decomposed matrices contain the most essential information of the input kernel, which ensures the proposed network to handle various blurs in a unified framework and generate high-quality deblurring results. Experimental results on benchmark datasets with noise and saturated pixels demonstrate that the proposed algorithm performs favorably against state-of-the-art methods.

In this paper, we propose a learning-based method to compose a video-story from a group of video clips that describe an activity or experience. We learn the coherence between video clips from real videos via the Recurrent Neural Network (RNN) that jointly incorporates the spatial-temporal semantics and motion dynamics to generate smooth and relevant compositions. We further rearrange the results generated by the RNN to make the overall video-story compatible with the storyline structure via a submodular ranking optimization process. Experimental results on the video-story dataset show that the proposed algorithm outperforms the state-of-the-art approach.

We propose a background subtraction algorithm using hierarchical superpixel segmentation, spanning trees and optical flow. First, we generate superpixel segmentation trees using a number of Gaussian Mixture Models (GMMs) by treating each GMM as one vertex to construct spanning trees. Next, we use the -smoother to enhance the spatial consistency on the spanning trees and estimate optical flow to extend the -smoother to the temporal domain. Experimental results on synthetic and real-world benchmark datasets show that the proposed algorithm performs favorably for background subtraction in videos against the state-of-the-art methods in spite of frequent and sudden changes of pixel values.

Images acquired in low-light conditions with handheld cameras are often blurry, so steady poses and long exposure time are required to alleviate this problem. Although significant advances have been made in image deblurring, state-of-the-art approaches often fail on low-light images, as a sufficient number of salient features cannot be extracted for blur kernel estimation. On the other hand, light streaks are common phenomena in low-light images that have not been extensively explored in existing approaches. In this work, we propose an algorithm that utilizes light streaks to facilitate deblurring low-light images. The light streaks, which commonly exist in the low-light blurry images, contain rich information regarding camera motion and blur kernels. A method is developed in this work to detect light streaks for kernel estimation. We introduce a non-linear blur model that explicitly takes light streaks and corresponding light sources into account, and pose them as constraints for estimating the blur kernel in an optimization framework. For practical applications, the proposed algorithm is extended to handle images undergoing non-uniform blur. Experimental results show that the proposed algorithm performs favorably against the state-of-the-art methods on deblurring real-world low-light images.

We present an effective blind image deblurring algorithm based on the dark channel prior. The motivation of this work is an interesting observation that the dark channel of blurred images is less sparse. While most patches in a clean image contain some dark pixels, this is not the case when they are averaged with neighboring ones by motion blur. This change in sparsity of the dark channel pixels is an inherent property of the motion blur process, which we prove mathematically and validate using image data. Enforcing sparsity of the dark channel thus helps blind deblurring in various scenarios such as natural, face, text, and low-illumination images. However, imposing sparsity of the dark channel introduces a non-convex non-linear optimization problem. In this work, we introduce a linear approximation to address this issue. Extensive experiments demonstrate that the proposed deblurring algorithm achieves the state-of-the-art results on natural images and performs favorably against methods designed for specific scenarios. In addition, we show that the proposed method can be applied to image dehazing.

In this paper, we propose a constrained graph labeling algorithm for visual tracking where nodes denote superpixels and edges encode the underlying spatial, temporal, and appearance fitness constraints. First, the spatial smoothness constraint, based on a transductive learning method, is enforced to leverage the latent manifold structure in feature space by investigating unlabeled superpixels in the current frame. Second, the appearance fitness constraint, which measures the probability of a superpixel being contained in the target region, is developed to incrementally induce a long-term appearance model. Third, the temporal smoothness constraint is proposed to construct a short-term appearance model of the target, which handles the drastic appearance change between consecutive frames. All these three constraints are incorporated in the proposed graph labeling algorithm such that induction and transduction, short- and long-term appearance models are combined, respectively. The foreground regions inferred by the proposed graph labeling method are used to guide the tracking process. Tracking results, in turn, facilitate more accurate online update by filtering out potential contaminated training samples. Both quantitative and qualitative evaluations on challenging tracking data sets show that the proposed constrained tracking algorithm performs favorably against the state-of-the-art methods.

Image retargeting has been applied to display images of any size via devices with various resolutions (e.g., cell phone, TV monitors). To fit an image with the target resolution, certain unimportant regions need to be deleted or distorted and the key problem is to determine the importance of each pixel. Existing methods predict pixel-wise importance in a bottom-up manner via eye fixation estimation or saliency detection. In contrast, the proposed algorithm estimates the pixel-wise importance based on a top-down criterion where the target image maintains the semantic meaning of the original image. To this end, several semantic components corresponding to foreground objects, action contexts, and background regions are extracted. The semantic component maps are integrated by a classification guided fusion network. Specifically, the deep network classifies the original image as object or scene-oriented, and fuses the semantic component maps according to classification results. The network output, referred to as the semantic collage with the same size as the original image, is then fed into any existing optimization method to generate the target image. Extensive experiments are carried out on the RetargetMe dataset and S-Retarget database developed in this work. Experimental results demonstrate the merits of the proposed algorithm over the state-of-the-art image retargeting methods.

Sparse coding has been applied to visual tracking and related vision problems with demonstrated success in recent years. Existing tracking methods based on local sparse coding sample patches from a target candidate and sparsely encode these using a dictionary consisting of patches sampled from target template images. The discriminative strength of existing methods based on local sparse coding is limited as spatial structure constraints among the template patches are not exploited. To address this problem, we propose a structure-aware local sparse coding algorithm, which encodes a target candidate using templates with both global and local sparsity constraints. For robust tracking, we show the local regions of a candidate region should be encoded only with the corresponding local regions of the target templates that are the most similar from the global view. Thus, a more precise and discriminative sparse representation is obtained to account for appearance changes. To alleviate the issues with tracking drifts, we design an effective template update scheme. Extensive experiments on challenging image sequences demonstrate the effectiveness of the proposed algorithm against numerous state-of-the-art methods.

In this paper, a novel spatial-temporal locality is proposed and unified via a discriminative dictionary learning framework for visual tracking. By exploring the strong local correlations between temporally obtained target and their spatially distributed nearby background neighbors, a spatial-temporal locality is obtained. The locality is formulated as a subspace model and exploited under a unified structure of discriminative dictionary learning with a subspace structure. Using the learned dictionary, the target and its background can be described and distinguished effectively through their sparse codes. As a result, the target is localized by integrating both the descriptive and the discriminative qualities. Extensive experiments on various challenging video sequences demonstrate the superior performance of proposed algorithm over the other state-of-the-art approaches.

Superpixel segmentation has been one of the most important tasks in computer vision. In practice, an object can be represented by a number of segments at finer levels with consistent details or included in a surrounding region at coarser levels. Thus, a superpixel segmentation hierarchy is of great importance for applications that require different levels of image details. However, there is no method that can generate all scales of superpixels accurately in real time. In this paper, we propose the superhierarchy algorithm which is able to generate multi-scale superpixels as accurately as the state-of-the-art methods but with one to two orders of magnitude speed-up. The proposed algorithm can be directly integrated with recent efficient edge detectors to significantly outperform the state-of-the-art methods in terms of segmentation accuracy. Quantitative and qualitative evaluations on a number of applications demonstrate that the proposed algorithm is accurate and efficient in generating a hierarchy of superpixels.

The success of the state-of-the-art deblurring methods mainly depends on the restoration of sharp edges in a coarse-to-fine kernel estimation process. In this paper, we propose to learn a deep convolutional neural network for extracting sharp edges from blurred images. Motivated by the success of the existing filtering-based deblurring methods, the proposed model consists of two stages: suppressing extraneous details and enhancing sharp edges. We show that the two-stage model simplifies the learning process and effectively restores sharp edges. Facilitated by the learned sharp edges, the proposed deblurring algorithm does not require any coarse-to-fine strategy or edge selection, thereby significantly simplifying kernel estimation and reducing computation load. Extensive experimental results on challenging blurry images demonstrate that the proposed algorithm performs favorably against the state-of-the-art methods on both synthetic and real-world images in terms of visual quality and run-time.

For image retrieval methods based on bag of visual words, much attention has been paid to enhancing the discriminative powers of the local features. Although retrieved images are usually similar to a query in minutiae, they may be significantly different from a semantic perspective, which can be effectively distinguished by convolutional neural networks (CNN). Such images should not be considered as relevant pairs. To tackle this problem, we propose to construct a dynamic match kernel by adaptively calculating the matching thresholds between query and candidate images based on the pairwise distance among deep CNN features. In contrast to the typical static match kernel which is independent to the global appearance of retrieved images, the dynamic one leverages the semantical similarity as a constraint for determining the matches. Accordingly, we propose a semantic-constrained retrieval framework by incorporating the dynamic match kernel, which focuses on matched patches between relevant images and filters out the ones for irrelevant pairs. Furthermore, we demonstrate that the proposed kernel complements recent methods, such as hamming embedding, multiple assignment, local descriptors aggregation, and graph-based re-ranking, while it outperforms the static one under various settings on off-the-shelf evaluation metrics. We also propose to evaluate the matched patches both quantitatively and qualitatively. Extensive experiments on five benchmark data sets and large-scale distractors validate the merits of the proposed method against the state-of-the-art methods for image retrieval.

In this paper, we propose a novel correlation particle filter (CPF) for robust visual tracking. Instead of a simple combination of a correlation filter and a particle filter, we exploit and complement the strength of each one. Compared with existing tracking methods based on correlation filters and particle filters, the proposed tracker has four major advantages: 1) it is robust to partial and total occlusions, and can recover from lost tracks by maintaining multiple hypotheses; 2) it can effectively handle large-scale variation via a particle sampling strategy; 3) it can efficiently maintain multiple modes in the posterior density using fewer particles than conventional particle filters, resulting in low computational cost; and 4) it can shepherd the sampled particles toward the modes of the target state distribution using a mixture of correlation filters, resulting in robust tracking performance. Extensive experimental results on challenging benchmark data sets demonstrate that the proposed CPF tracking algorithm performs favorably against the state-of-the-art methods.