Despite recent progress, video generative models still struggle to generate delicate human actions (e.g., gymnastics), particularly when they are required to start from a user-provided reference image. In this paper, we explore the task of learning to animate images into videos that portray delicate human actions using a small number of videos — 16 or fewer — which reduces the need for extensive data collection and enhances practicality for real-world applications. Learning generalizable motion patterns that smoothly transition from user-provided reference images in such a few-shot setting is highly challenging. We propose FLASH (Few-shot Learning to Animate and Steer Humans), which enhances generalization of motion by training the model to reconstruct a video using the motion features and cross-frame correspondences extracted from another video with the same motion but different appearance. This encourages the learning of transferable motion and mitigates overfitting to the appearance in limited training data. Additionally, FLASH extends the decoder with additional layers to propagate details from the reference image to generated frames, improving transition smoothness. Human judges significantly favor FLASH, with 65.78% of 488 responses prefer FLASH over baselines. We strongly recommend watching the videos on the webpage1, as motion artifacts are hard to notice from images.

Multimodal Large Language Models (MLLMs) have shown exceptional capabilities in vision-language tasks. However, effectively integrating image segmentation into these models remains a significant challenge.In this work, we propose a novel text-as-mask paradigm that casts image segmentation as a text generation problem, eliminating the need for additional decoders and significantly simplifying the segmentation process. Our key innovation is semantic descriptors, a new textual representation of segmentation masks where each image patch is mapped to its corresponding text label. We first introduce image-wise semantic descriptors, a patch-aligned textual representation of segmentation masks that integrates naturally into the language modeling pipeline. To enhance efficiency, we introduce the Row-wise Run-Length Encoding (R-RLE), which compresses redundant text sequences, reducing the length of semantic descriptorsby 74% and accelerating inference by $3\times$3×, without compromising performance. Building upon this, our initial framework Text4Segachieves strong segmentation performance across a wide range of vision tasks. To further improve granularity and compactness, we propose box-wise semantic descriptors, which localizes regions of interest using bounding boxes and represents region masks via structured mask tokens called semantic bricks. This leads to our refined model, Text4Seg++, which formulates segmentation as a next-brick prediction task, combining precision, scalability, and generative efficiency. Comprehensive experiments on natural and remote sensing datasets show that Text4Seg++consistently outperforms state-of-the-art models across diverse benchmarks without any task-specific fine-tuning, while remaining compatible with existing MLLM backbones. Our work highlights the effectiveness, scalability, and generalizability of text-driven image segmentation within the MLLM framework.

Video Temporal Grounding (VTG) aims to precisely identify video event segments in response to textual queries. The outputs of VTG tasks manifest as sequences of events, each defined by precise timestamps, saliency scores, and textual descriptions. Despite recent advances, a fundamental limitation persists in existing Video Large Language Models (Video-LLMs): they process all task tokens through identical and static pathways, failing to recognize that temporal localization, saliency assessment, and textual generation represent fundamentally distinct tasks requiring specialized processing. To address this, we introduce TimeExpert, a Mixture-of-Experts (MoE)-based Video-LLM that effectively decomposes VTG tasks by dynamically routing task-specific tokens (e.g., timestamps, saliency scores) to specialized experts, with increased computational efficiency. Our design choices enable precise handling of each subtask, leading to improved event modeling across diverse VTG applications. Extensive experiments demonstrate that TimeExpert consistently achieves state-of-the-art performance on various VTG tasks such as Dense Video Captioning, Moment Retrieval, and Video Highlight Detection.

Leveraging text, images, structure maps, or motion trajectories as conditional guidance, diffusion models have achieved great success in automated and high-quality video generation. However, generating smooth and rational transition videos given the first and last video frames as well as descriptive text prompts is far underexplored. We present VTG, a Versatile Transition video Generation framework that can generate smooth, high-fidelity, and semantically coherent video transitions. VTG introduces interpolation-based initialization that helps preserve object identity and handle abrupt content changes effectively. In addition, it incorporates dual-directional motion fine-tuning and representation alignment regularization to mitigate the limitations of pre-trained image-to-video diffusion models in motion smoothness and generation fidelity, respectively. To evaluate VTG and facilitate future studies on unified transition generation, we collected TransitBench, a comprehensive benchmark for transition generation covering two representative transition tasks: concept blending and scene transition. Extensive experiments show that VTG achieves superior transition performance consistently across all four tasks.

While Multimodal Large Language Models (MLLMs) have advanced GUI navigation agents, current approaches face limitations in cross-domain generalization and effective history utilization. We present a reasoning-enhanced framework that systematically integrates structured reasoning, action prediction, and history summarization. The structured reasoning component generates coherent Chain-of-Thought analyses combining progress estimation and decision reasoning, which inform both immediate action predictions and compact history summaries for future steps. Based on this framework, we train a GUI agent, GUI-Rise, through supervised fine-tuning on pseudo-labeled trajectories and reinforcement learning with Group Relative Policy Optimization (GRPO). This framework employs specialized rewards, including a history-aware objective, directly linking summary quality to subsequent action performance. Comprehensive evaluations on standard benchmarks demonstrate state-of-the-art results under identical training data conditions, with particularly strong performance in out-of-domain scenarios. These findings validate our framework's ability to maintain robust reasoning and generalization across diverse GUI navigation tasks.

Person search is a challenging task in computer vision and multimedia understanding, which aims at localizing and identifying target individuals in realistic scenes. State-of-the-art models achieve remarkable success but suffer from overloaded computation and inefficient inference, making them impractical in most real-world applications. A promising approach to tackle this dilemma is to compress person search models with knowledge distillation (KD). Previous KD-based person search methods typically distill the knowledge from the re-identification (re-id) branch, completely overlooking the useful knowledge from the detection branch. In addition, we elucidate that the imbalance between person and background regions in feature maps has a negative impact on the distillation process. To this end, we propose a novel KD-based approach, namely Disaggregation Distillation for Person Search (DDPS), which disaggregates the distillation process and feature maps, respectively. Firstly, the distillation process is disaggregated into two task-oriented sub-processes,i.e., detection distillation and re-id distillation, to help the student learn both accurate localization capability and discriminative person embeddings. Secondly, we disaggregate each feature map into person and background regions, and distill these two regions independently to alleviate the imbalance problem. More concretely, three types of distillation modules,i.e., logit distillation (LD), correlation distillation (CD), and disaggregation feature distillation (DFD), are particularly designed to transfer comprehensive information from the teacher to the student. Note that such a simple yet effective distillation scheme can be readily applied to both homogeneous and heterogeneous teacher-student combinations. We conduct extensive experiments on two person search benchmarks, where the results demonstrate that, surprisingly, our DDPS enables the student model to surpass the performance of the corresponding teacher model, even achieving comparable results with general person search models.

We present DIRECT-3D, a diffusion-based 3D generative model for creating high-quality 3D assets (represented by Neural Radiance Fields) from text prompts. Unlike recent 3D generative models that rely on clean and well-aligned 3D data, limiting them to single or few-class generation, our model is directly trained on extensive noisy and unaligned ‘in-the-wild’ 3D assets, mitigating the key challenge (i.e., data scarcity) in large-scale 3D generation. In particular, DIRECT-3D is a tri-plane diffusion model that integrates two innovations: 1) A novel learning framework where noisy data are filtered and aligned automatically during the training process. Specifically, after an initial warm-up phase using a small set of clean data, an iterative optimization is introduced in the diffusion process to explicitly estimate the 3D pose of objects and select beneficial data based on conditional density. 2) An efficient 3D representation that is achieved by disentangling object geometry and color features with two separate conditional diffusion models that are optimized hierarchically. Given a prompt input, our model generates high-quality, high-resolution, realistic, and complex 3D objects with accurate geometric details in seconds. We achieve state-of-the-art performance in both single-class generation and text-to-3D generation. We also demonstrate that DIRECT-3D can serve as a useful 3D geometric prior of objects, for example to alleviate the well-known Janus problem in 2D-lifting methods such as DreamFusion. The code and models are available for research purposes at: https://github.com/qihao067/direct3d.

Accurate and controllable image editing is a challenging task that has attracted significant attention recently. Notably, DRAGGAN developed by Pan et al. (2023) [33] is an interactive point-based image editing framework that achieves impressive editing results with pixel-level precision. However, due to its reliance on generative adversarial networks (GANs), its generality is limited by the capacity of pretrained GAN models. In this work, we extend this editing framework to diffusion models and propose a novel approach Dragdiffusion. By harnessing large-scale pretrained diffusion models, we greatly enhance the applicability of interactive point-based editing on both real and diffusion-generated images. Unlike other diffusion-based editing methods that provide guidance on diffusion latents of multiple time steps, our approach achieves efficient yet accurate spatial control by optimizing the latent of only one time step. This novel design is motivated by our observations that UNet features at a specific time step provides sufficient semantic and geometric information to support the drag-based editing. Moreover, we introduce two additional techniques, namely identity-preserving fine-tuning and reference-latent-control, to further preserve the identity of the original image. Lastly, we present a challenging benchmark dataset called DRAGBENCH─ the first benchmark to evaluate the performance of interactive point-based image editing methods. Experiments across a wide range of challenging cases (e.g., images with multiple objects, diverse object categories, various styles, etc.) demonstrate the versatility and generality of Dragdiffusion. Code and the Dragbench dataset: https://github.com/Yujun-Shi/DragDiffusion.

We present GLEE in this work, an object-level foundation model for locating and identifying objects in images and videos. Through a unified framework, GLEE accomplishes detection, segmentation, tracking, grounding, and identification of arbitrary objects in the open world scenario for various object perception tasks. Adopting a cohesive learning strategy, GLEE acquires knowledge from diverse data sources with varying supervision levels to formu-late general object representations, excelling in zero-shot transfer to new data and tasks. Specifically, we employ an image encoder, text encoder, and visual prompter to handle multimodal inputs, enabling to simultaneously solve various object-centric downstream tasks while maintaining state-of-the-art performance. Demonstrated through extensive training on over five million images from diverse benchmarks, GLEE exhibits remarkable versatility and improved generalization performance, efficiently tack-ling downstream tasks without the need for task-specific adaptation. By integrating large volumes of automatically labeled data, we further enhance its zero-shot generalization capabilities. Additionally, GLEE is capable of being integrated into Large Language Models, serving as a foun-dational model to provide universal object-level information for multimodal tasks. We hope that the versatility and universality of our method will mark a significant step in the development of efficient visual foundation models for AGI systems. The models and code are released at https://github.com/FoundationVision/GLEE.

Text-guided diffusion models (TDMs) are widely applied but can fail unexpectedly. Common failures include: _(i)_ natural-looking text prompts generating images with the wrong content, or _(ii)_ different random samples of the latent variables that generate vastly different, and even unrelated, outputs despite being conditioned on the same text prompt. In this work, we aim to study and understand the failure modes of TDMs in more detail. To achieve this, we propose SAGE, the first adversarial search method on TDMs that systematically explores the discrete prompt space and the high-dimensional latent space, to automatically discover undesirable behaviors and failure cases in image generation. We use image classifiers as surrogate loss functions during searching, and employ human inspections to validate the identified failures. For the first time, our method enables efficient exploration of both the discrete and intricate human language space and the challenging latent space, overcoming the gradient vanishing problem. Then, we demonstrate the effectiveness of SAGE on five widely used generative models and reveal four typical failure modes that have not been systematically studied before: (1) We find a variety of natural text prompts that generate images failing to capture the semantics of input texts. We further discuss the underlying causes and potential solutions based on the results. (2) We find regions in the latent space that lead to distorted images independent of the text prompt, suggesting that parts of the latent space are not well-structured. (3) We also find latent samples that result in natural-looking images unrelated to the text prompt, implying a possible misalignment between the latent and prompt spaces. (4) By appending a single adversarial token embedding to any input prompts, we can generate a variety of specified target objects, with minimal impact on CLIP scores, demonstrating the fragility of language representations.

Open-world instance-level scene understanding aims to locate and recognize unseen object categories that are not present in the annotated dataset. This task is challenging because the model needs to both localize novel 3D objects and infer their semantic categories. A key factor for the recent progress in 2D open-world perception is the availability of large-scale image-text pairs from the Internet, which cover a wide range of vocabulary concepts. However, this success is hard to replicate in 3D scenarios due to the scarcity of 3D-text pairs. To address this challenge, we propose to harness pre-trained vision-language (VL) foundation models that encode extensive knowledge from image-text pairs to generate captions for multi-view images of 3D scenes. This allows us to establish explicit associations between 3D shapes and semantic-rich captions. Moreover, to enhance the fine-grained visual-semantic representation learning from captions for object-level categorization, we design hierarchical point-caption association methods to learn semantic-aware embeddings that exploit the 3D geometry between 3D points and multi-view images. In addition, to tackle the localization challenge for novel classes in the open-world setting, we develop debiased instance localization, which involves training object grouping modules on unlabeled data using instance-level pseudo supervision. This significantly improves the generalization capabilities of instance grouping and, thus, the ability to accurately locate novel objects. We conduct extensive experiments on 3D semantic, instance, and panoptic segmentation tasks, covering indoor and outdoor scenes across three datasets. Our method outperforms baseline methods by a significant margin in semantic segmentation (e.g. 34.5% ∼ 65.3%), instance segmentation (e.g. 21.8% ∼ 54.0%), and panoptic segmentation (e.g. 14.7% ∼ 43.3%). Code will be available.

There are two mainstream approaches for object detection: top-down and bottom-up. The state-of-the-art approaches are mainly top-down methods. In this paper, we demonstrate that bottom-up approaches show competitive performance compared with top-down approaches and have higher recall rates. Our approach, named CenterNet, detects each object as a triplet of keypoints (top-left and bottom-right corners and the center keypoint). We first group the corners according to some designed cues and confirm the object locations based on the center keypoints. The corner keypoints allow the approach to detect objects of various scales and shapes and the center keypoint reduces the confusion introduced by a large number of false-positive proposals. Our approach is an anchor-free detector because it does not need to define explicit anchor boxes. We adapt our approach to backbones with different structures, including 'hourglass'-like networks and 'pyramid'-like networks, which detect objects in single-resolution and multi-resolution feature maps, respectively. On the MS-COCO dataset, CenterNet with Res2Net-101 and Swin-Transformer achieve average precisions (APs) of 53.7% and 57.1%, respectively, outperforming all existing bottom-up detectors and achieving state-of-the-art performance. We also design a real-time CenterNet model, which achieves a good trade-off between accuracy and speed, with an AP of 43.6% at 30.5 frames per second (FPS).

Bridge detection in remote sensing images (RSIs) plays a crucial role in various applications, but it poses unique challenges compared to the detection of other objects. In RSIs, bridges exhibit considerable variations in terms of their spatial scales and aspect ratios. Therefore, to ensure the visibility and integrity of bridges, it is essential to perform holistic bridge detection in large-size very-high-resolution (VHR) RSIs. However, the lack of datasets with large-size VHR RSIs limits the deep learning algorithms' performance on bridge detection. Due to the limitation of GPU memory in tackling large-size images, deep learning-based object detection methods commonly adopt the cropping strategy, which inevitably results in label fragmentation and discontinuous prediction. To ameliorate the scarcity of datasets, this paper proposes a large-scale dataset named GLH-Bridge comprising 6,000 VHR RSIs sampled from diverse geographic locations across the globe. These images encompass a wide range of sizes, varying from 2,048 × 2,048 to 16,384 × 16,384 pixels, and collectively feature 59,737 bridges. These bridges span diverse backgrounds, and each of them has been manually annotated, using both an oriented bounding box (OBB) and a horizontal bounding box (HBB). Furthermore, we present an efficient network for holistic bridge detection (HBD-Net) in large-size RSIs. The HBD-Net presents a separate detector-based feature fusion (SDFF) architecture and is optimized via a shape-sensitive sample re-weighting (SSRW) strategy. The SDFF architecture performs inter-layer feature fusion (IFF) to incorporate multi-scale context in the dynamic image pyramid (DIP) of the large-size image, and the SSRW strategy is employed to ensure an equitable balance in the regression weight of bridges with various aspect ratios. Based on the proposed GLH-Bridge dataset, we establish a bridge detection benchmark including the OBB and HBB tasks, and validate the effectiveness of the proposed HBD-Net. Additionally, cross-dataset generalization experiments on two publicly available datasets illustrate the strong generalization capability of the GLH-Bridge dataset.

Federated learning aims to train models collaboratively across different clients without sharing data for privacy considerations. However, one major challenge for this learning paradigm is thedata heterogeneityproblem, which refers to the discrepancies between the local data distributions among various clients. To tackle this problem, we first study how data heterogeneity affects the representations of the globally aggregated models. Interestingly, we find that heterogeneous data results in the global model suffering from severedimensional collapse, in which representations tend to reside in a lower-dimensional space instead of the ambient space. This dimensional collapse phenomenon severely curtails the expressive power of models, leading to significant degradation in the performance. Next, via experiments, we make more observations and posit two reasons that result in this phenomenon: 1) dimensional collapse on local models; 2) the operation of global averaging on local model parameters. In addition, we theoretically analyze the gradient flow dynamics to shed light on how data heterogeneity result in dimensional collapse. To remedy this problem caused by the data heterogeneity, we proposeFedDecorr, a novel method that can effectively mitigate dimensional collapse in federated learning. Specifically,FedDecorrapplies a regularization term during local training that encourages different dimensions of representations to be uncorrelated.FedDecorr, which is implementation-friendly and computationally-efficient, yields consistent improvements over various baselines on five standard benchmark datasets including CIFAR10, CIFAR100, TinyImageNet, Office-Caltech10, and DomainNet.

Open-vocabulary scene understanding aims to localize and recognize unseen categories beyond the annotated label space. The recent breakthrough of 2D open-vocabulary perception is largely driven by Internet-scale paired image-text data with rich vocabulary concepts. However, this success cannot be directly transferred to 3D scenarios due to the inaccessibility of large-scale 3D-text pairs. To this end, we propose to distill knowledge encoded in pretrained vision-language (VL) foundation models through captioning multi-view images from 3D, which allows explicitly associating 3D and semantic-rich captions. Further, to foster coarse-to-fine visual-semantic representation learning from captions, we design hierarchical 3D-caption pairs, leveraging geometric constraints between 3D scenes and multi-view images. Finally, by employing contrastive learning, the model learns language-aware embeddings that connect 3D and text for open-vocabulary tasks. Our method not only remarkably outperforms baseline methods by 25.8% ~ 44.7% hIoU and 14.5% ~ 50.4% hAP50in open-vocabulary semantic and instance segmentation, but also shows robust transferability on challenging zero-shot domain transfer tasks. See the project website at https://dingry.github.io/projects/PLA.

Despite significant efforts, cutting-edge video segmentation methods still remain sensitive to occlusion and rapid movement, due to their reliance on the appearance of objects in the form of object embeddings, which are vulnerable to these disturbances. A common solution is to use optical flow to provide motion information, but essentially it only considers pixel-level motion, which still relies on appearance similarity and hence is often inaccurate under occlusion and fast movement. In this work, we study the instance-level motion and present InstMove, which stands for Instance Motion for Object-centric Video Segmentation. In comparison to pixel-wise motion, Inst-Move mainly relies on instance-level motion information that is free from image feature embeddings, and features physical interpretations, making it more accurate and robust toward occlusion and fast-moving objects. To better fit in with the video segmentation tasks, InstMove uses instance masks to model the physical presence of an object and learns the dynamic model through a memory network to predict its position and shape in the next frame. With only a few lines of code, InstMove can be integrated into current SOTA methods for three different video segmentation tasks and boost their performance. Specifically, we improve the previous arts by 1.5 AP on OVIS dataset, which features heavy occlusions, and 4.9 AP on YouTube VIS-Long dataset, which mainly contains fast moving objects. These results suggest that instance-level motion is robust and accurate, and hence serving as a powerful solution in complex scenarios for object-centric video segmentation.

Video object segmentation (VOS) aims at segmenting a particular object throughout the entire video clip sequence. The state-of-the-art VOS methods have achieved excellent performance (e.g., 90+% $\mathcal{J}$ & $\mathcal{F}$) on existing datasets. However, since the target objects in these existing datasets are usually relatively salient, dominant, and isolated, VOS under complex scenes has rarely been studied. To revisit VOS and make it more applicable in the real world, we collect a new VOS dataset called coMplex video Object SEgmentation (MOSE) to study the tracking and segmenting objects in complex scenarios. MOSE contains 2,149 video clips and 5,200 objects from 36 categories, with 431,725 high-quality object segmentation masks. The most notable feature of MOSE dataset is complex scenes with crowded and occluded objects. The target objects in the videos are commonly occluded by others and disappear in some frames. To analyze the proposed MOSE dataset, we benchmark 18 existing VOS methods under 4 different settings on the proposed MOSE dataset and conduct comprehensive comparisons. The experiments show that current VOS algorithms cannot well perceive objects in complex scenes. For example, under the semi-supervised VOS setting, the highest $\mathcal{J}$ & $\mathcal{F}$ by existing state-of-the-art VOS methods is only 59.4% on MOSE, much lower than their ∼90% $\mathcal{J}$ & $\mathcal{F}$ performance on DAVIS. The results reveal that although excellent performance has been achieved on existing benchmarks, there are unresolved challenges under complex scenes and more efforts are desired to explore these challenges in the future.

Previous works have shown that increasing the window size for Transformer-based image super-resolution models (e.g., SwinIR) can significantly improve the model performance but the computation overhead is also considerable. In this paper, we present SRFormer, a simple but novel method that can enjoy the benefit of large window self-attention but introduces even less computational burden. The core of our SRFormer is the permuted self-attention (PSA), which strikes an appropriate balance between the channel and spatial information for self-attention. Our PSA is simple and can be easily applied to existing super-resolution networks based on window self-attention. Without any bells and whistles, we show that our SRFormer achieves a 33.86dB PSNR score on the Urban100 dataset, which is 0.46dB higher than that of SwinIR but uses fewer parameters and computations. We hope our simple and effective approach can serve as a useful tool for future research in super-resolution model design. Our code is available at https://github.com/HVision-NKU/SRFormer.

Unsupervised domain adaptation (UDA) has been widely applied in improving model generalization on unlabeled target data. However, accurately selecting the best UDA model for the target domain is challenging due to the absence of labeled target data and domain distribution shifts. Traditional model selection approaches involve training extra models with source data to estimate the target validation risk. Recent studies propose practical methods that are based on measuring various properties of model predictions on target data. Although effective for some UDA models, these methods often lack stability and may lead to poor selections for other UDA models. In this paper, we present MixVal, an innovative model selection method that operates solely with unlabeled target data during inference. MixVal leverages mixed target samples with pseudo labels to directly probe the learned target structure by each UDA model. Specifically, MixVal employs two distinct types of probes: the intra-cluster mixed samples for evaluating neighborhood density and the inter-cluster mixed samples for investigating the classification boundary. With this comprehensive probing strategy, MixVal elegantly combines the strengths of two state-of-the-art model selection methods, Entropy and SND. We extensively evaluate MixVal on 11 UDA methods across 4 adaptation settings, including classification and segmentation tasks. Experimental results consistently demonstrate that MixVal achieves state-of-the-art performance and maintains exceptional stability in model selection. Code is available at \url{https://github.com/LHXXHB/MixVal}.

The goal of this special section is to provide a platform to summarize what we have achieved and where we are moving toward this methodological research direction. It gathers the latest advances in learning with graph-structured data in computer vision, as well as interdisciplinary efforts on graph-based pattern analysis in sociology, physics, chemistry, finance, biology, etc.

The computational complexity of transformers limits it to be widely deployed onto frameworks for visual recognition. Recent work [9] significantly accelerates the network processing speed by reducing the resolution at the beginning of the network, however, it is still hard to be directly generalized onto other downstream tasks e.g. object detection and segmentation like CNN. In this paper, we present a transformer-based architecture retaining both the local and global interactions within the network, and can be transferable to other downstream tasks. The proposed architecture reforms the original full spatial self-attention into pixel-wise local attention and patch-wise global attention. Such factorization saves the computational cost while retaining the information of different granularities, which helps generate multi-scale features required by different tasks. By exploiting the factorized attention, we construct a Separable Transformer (SeT) for visual modeling. Experimental results show that SeT outperforms the previous state-of-the-art transformer-based approaches and its CNN counterparts on three major tasks including image classification, object detection and instance segmentation.

Leveraging the advances of natural language processing, most recent scene text recognizers adopt an encoder-decoder architecture where text images are first converted to representative features and then a sequence of characters via 'sequential decoding'. However, scene text images suffer from rich noises of different sources such as complex background and geometric distortions which often confuse the decoder and lead to incorrect alignment of visual features at noisy decoding time steps. This paper presents I2C2W, a novel scene text recognition technique that is tolerant to geometric and photometric degradation by decomposing scene text recognition into two inter-connected tasks. The first task focuses on image-to-character (I2C) mapping which detects a set of character candidates from images based on different alignments of visual features in an non-sequential way. The second task tackles character-to-word (C2W) mapping which recognizes scene text by decoding words from the detected character candidates. The direct learning from character semantics (instead of noisy image features) corrects falsely detected character candidates effectively which improves the final text recognition accuracy greatly. Extensive experiments over nine public datasets show that the proposed I2C2W outperforms the state-of-the-art by large margins for challenging scene text datasets with various curvature and perspective distortions. It also achieves very competitive recognition performance over multiple normal scene text datasets.

This article presents Holistically-Attracted Wireframe Parsing (HAWP), a method for geometric analysis of 2D images containing wireframes formed by line segments and junctions. HAWP utilizes a parsimonious Holistic Attraction (HAT) field representation that encodes line segments using a closed-form 4D geometric vector field. The proposed HAWP consists of three sequential components empowered by end-to-end and HAT-driven designs: 1) generating a dense set of line segments from HAT fields and endpoint proposals from heatmaps, 2) binding the dense line segments to sparse endpoint proposals to produce initial wireframes, and 3) filtering false positive proposals through a novel endpoint-decoupled line-of-interest aligning (EPD LOIAlign) module that captures the co-occurrence between endpoint proposals and HAT fields for better verification. Thanks to our novel designs, HAWPv2 shows strong performance in fully supervised learning, while HAWPv3 excels in self-supervised learning, achieving superior repeatability scores and efficient training (24 GPU hours on a single GPU). Furthermore, HAWPv3 exhibits a promising potential for wireframe parsing in out-of-distribution images without providing ground truth labels of wireframes.

Data-driven discovery of image-derived phenotypes (IDPs) from large-scale multimodal brain imaging data has enormous potential for neuroscientific and clinical research by linking IDPs to subjects' demographic, behavioural, clinical and cognitive measures (i.e., non-imaging derived phenotypes or nIDPs). However, current approaches are primarily based on unsupervised approaches, without the use of information in nIDPs. In this paper, we proposed a semi-supervised, multimodal, and multi-task fusion approach, termed SuperBigFLICA, for IDP discovery, which simultaneously integrates information from multiple imaging modalities as well as multiple nIDPs. SuperBigFLICA is computationally efficient and largely avoids the need for parameter tuning. Using the UK Biobank brain imaging dataset with around 40,000 subjects and 47 modalities, along with more than 17,000 nIDPs, we showed that SuperBigFLICA enhances the prediction power of nIDPs, benchmarked against IDPs derived by conventional expert-knowledge and unsupervised-learning approaches (with average nIDP prediction accuracy improvements of up to 46%). It also enables the learning of generic imaging features that can predict new nIDPs. Further empirical analysis of the SuperBigFLICA algorithm demonstrates its robustness in different prediction tasks and the ability to derive biologically meaningful IDPs in predicting health outcomes and cognitive nIDPs, such as fluid intelligence and hypertension.

Many video understanding tasks require analyzing multishot videos, but existing datasets for video object segmentation (VOS) only consider single-shot videos. To address this challenge, we collected a new dataset-YouMVaS-of 200 popular YouTube videos spanning ten genres, where each video is on average five minutes long and with 75 shots. We selected recurring actors and annotated 431K segmentation masks at a frame rate of six, exceeding previous datasets in average video duration, object variation, and narrative structure complexity. We incorporated good practices of model architecture design, memory management, and multi-shot tracking into an existing video segmentation method to build competitive baseline methods. Through error analysis, we found that these baselines still fail to cope with cross-shot appearance variation on our YouMVOS dataset. Thus, our dataset poses new challenges in multi-shot segmentation towards better video analysis. Data, code, and pre-trained models are available at https://donglaiw.github.io/proj/youMVOS

Mixup-based augmentation has been found to be effective for generalizing models during training, especially for Vision Transformers (ViTs) since they can easily overfit. However, previous mixup-based methods have an underlying prior knowledge that the linearly interpolated ratio of targets should be kept the same as the ratio proposed in input interpolation. This may lead to a strange phenomenon that sometimes there is no valid object in the mixed image due to the random process in augmentation but there is still response in the label space. To bridge such gap between the input and label spaces, we propose TransMix, which mixes labels based on the attention maps of Vision Transformers. The confidence of the label will be larger if the corresponding input image is weighted higher by the attention map. TransMix is embarrassingly simple and can be implemented in just a few lines of code without introducing any extra parameters and FLOPs to ViT-based models. Experimental results show that our method can consistently improve various ViT-based models at scales on ImageNet classification. After pre-trained with TransMix on ImageNet, the ViT-based models also demonstrate better transferability to semantic segmentation, object detection and instance segmentation. TransMix also exhibits to be more robust when evaluating on 4 different benchmarks. Code is publicly available at https://github.com/Beckschen/TransMix.

Large-scale pre-training has been proven to be crucial for various computer vision tasks. However, with the increase of pre-training data amount, model architecture amount, and the private/inaccessible data, it is not very efficient or possible to pre-train all the model architectures on large-scale datasets. In this work, we investigate an alternative strategy for pre-training, namely Knowledge Distillation as Efficient Pre-training (KDEP), aiming to efficiently transfer the learned feature representation from existing pre-trained models to new student models for future downstream tasks. We observe that existing Knowledge Distillation (KD) methods are unsuitable towards pre-training since they normally distill the logits that are going to be discarded when transferred to downstream tasks. To resolve this problem, we propose a feature-based KD method with non-parametric feature dimension aligning. Notably, our method performs comparably with supervised pre-training counterparts in 3 downstream tasks and 9 downstream datasets requiring 10× less data and 5× less pre-training time. Code is available at https://github.com/CVMI-Lab/KDEP.

Temporal action detection (TAD) is an important yet challenging task in video understanding. It aims to simultaneously predict the semantic label and the temporal interval of every action instance in an untrimmed video. Rather than end-to-end learning, most existing methods adopt a head-only learning paradigm, where the video encoder is pre-trained for action classification, and only the detection head upon the encoder is optimized for TAD. The effect of end-to-end learning is not systematically evaluated. Besides, there lacks an in-depth study on the efficiency-accuracy trade-off in end-to-end TAD. In this paper, we present an empirical study of end-to-end temporal action detection. We validate the advantage of end-to-end learning over head-only learning and observe up to 11% performance improvement. Besides, we study the effects of multiple design choices that affect the TAD performance and speed, including detection head, video encoder, and resolution of input videos. Based on the findings, we build a mid-resolution baseline detector, which achieves the state-of-the-art performance of end-to-end methods while running more than 4×faster. We hope that this paper can serve as a guide for end-to-end learning and inspire future research in this field. Code and models are available at https://github.com/xlliu7/E2E-TAD.

Class Incremental Learning (CIL) aims at learning a classifier in a phase-by-phase manner, in which only data of a subset of the classes are provided at each phase. Previous works mainly focus on mitigating forgetting in phases after the initial one. However, we find that improving CIL at its initial phase is also a promising direction. Specifically, we experimentally show that directly encouraging CIL Learner at the initial phase to output similar representations as the model jointly trained on all classes can greatly boost the CIL performance. Motivated by this, we study the differ-ence between a naively-trained initial-phase model and the oracle model. Specifically, since one major difference be-tween these two models is the number of training classes, we investigate how such difference affects the model rep-resentations. We find that, with fewer training classes, the data representations of each class lie in a long and narrow region; with more training classes, the representations of each class scatter more uniformly. Inspired by this obser-vation, we propose Class-wise Decorrelation (CwD) that ef-fectively regularizes representations of each class to scatter more uniformly, thus mimicking the model jointly trained with all classes (i.e., the oracle model). Our CwD is simple to implement and easy to plug into existing methods. Ex-tensive experiments on various benchmark datasets show that CwD consistently and significantly improves the per-formance of existing state-of-the-art methods by around 1% to 3%. Code: https://github.com/Yujun-Shi/CwD.

A typical pipeline for multi-object tracking (MOT) is to use a detector for object localization, and following re-identification (re-ID)for object association. This pipeline is partially motivated by recent progress in both object detection and re- ID, and partially motivated by biases in existing tracking datasets, where most objects tend to have distin-guishing appearance and re-ID models are sufficient for es-tablishing associations. In response to such bias, we would like to re-emphasize that methods for multi-object tracking should also work when object appearance is not sufficiently discriminative. To this end, we propose a large-scale dataset for multi-human tracking, where humans have sim-ilar appearance, diverse motion and extreme articulation. As the dataset contains mostly group dancing videos, we name it “DanceTrack”. We expect DanceTrack to provide a better platform to develop more MOT algorithms that rely less on visual discrimination and depend more on motion analysis. We benchmark several state-of-the-art trackers on our dataset and observe a significant performance drop on DanceTrack when compared against existing benchmarks. The dataset, project code and competition is released at: https://github.com/DanceTrack.

State-of-the-art document dewarping techniques learn to predict 3-dimensional information of documents which are prone to errors while dealing with documents with irregular distortions or large variations in depth. This paper presents FDRNet, a Fourier Document Restoration Network that can restore documents with different distortions and improve document recognition in a reliable and simpler manner. FDRNet focuses on high-frequency components in the Fourier space that capture most structural information but are largely free of degradation in appearance. It dewarps documents by a flexible Thin-Plate Spline transformation which can handle various deformations effectively without requiring deformation annotations in training. These features allow FDRNet to learn from a small amount of simply labeled training images, and the learned model can dewarp documents with complex geometric distortion and recognize the restored texts accurately. To facilitate document restoration research, we create a benchmark dataset consisting of over one thousand camera documents with different types of geometric and photometric distortion. Extensive experiments show that FDRNet outperforms the state-of-the-art by large margins on both dewarping and text recognition tasks. In addition, FDRNet requires a small amount of simply labeled training data and is easy to deploy. The proposed dataset is available at https://sg-vilab.github.io/event/warpdoc/.

Neural Radiance Fields (NeRF) has demonstrated very impressive performance in novel view synthesis via implicitly modelling 3D representations from multi-view 2D images. However, most existing studies train NeRF models with either reasonable camera pose initialization or manually-crafted camera pose distributions which are often unavailable or hard to acquire in various real-world data. We design VMRF, an innovative view matching NeRF that enables effective NeRF training without requiring prior knowledge in camera poses or camera pose distributions. VMRF introduces a view matching scheme, which exploits unbalanced optimal transport to produce a feature transport plan for mapping a rendered image with randomly initialized camera pose to the corresponding real image. With the feature transport plan as the guidance, a novel pose calibration technique is designed which rectifies the initially randomized camera poses by predicting relative pose transformations between the pair of rendered and real images. Extensive experiments over a number of synthetic and real datasets show that the proposed VMRF outperforms the state-of-the-art qualitatively and quantitatively by large margins.

Abstract Can our video understanding systems perceive objects when a heavy occlusion exists in a scene? To answer this question, we collect a large-scale dataset called OVIS for occluded video instance segmentation, that is, to simultaneously detect, segment, and track instances in occluded scenes. OVIS consists of 296k high-quality instance masks from 25 semantic categories, where object occlusions usually occur. While our human vision systems can understand those occluded instances by contextual reasoning and association, our experiments suggest that current video understanding systems cannot. On the OVIS dataset, the highest AP achieved by state-of-the-art algorithms is only 16.3, which reveals that we are still at a nascent stage for understanding objects, instances, and videos in a real-world scenario. We also present a simple plug-and-play module that performs temporal feature calibration to complement missing object cues caused by occlusion. Built upon MaskTrack R-CNN and SipMask, we obtain a remarkable AP improvement on the OVIS dataset. The OVIS dataset and project code are available at http://songbai.site/ovis .

Temporal action detection (TAD) aims to determine the semantic label and the temporal interval of every action instance in an untrimmed video. It is a fundamental and challenging task in video understanding. Previous methods tackle this task with complicated pipelines. They often need to train multiple networks and involve hand-designed operations, such as non-maximal suppression and anchor generation, which limit the flexibility and prevent end-to-end learning. In this paper, we propose an end-to-end Transformer-based method for TAD, termed TadTR. Given a small set of learnable embeddings called action queries, TadTR adaptively extracts temporal context information from the video for each query and directly predicts action instances with the context. To adapt Transformer to TAD, we propose three improvements to enhance its locality awareness. The core is a temporal deformable attention module that selectively attends to a sparse set of key snippets in a video. A segment refinement mechanism and an actionness regression head are designed to refine the boundaries and confidence of the predicted instances, respectively. With such a simple pipeline, TadTR requires lower computation cost than previous detectors, while preserving remarkable performance. As a self-contained detector, it achieves state-of-the-art performance on THUMOS14 (56.7% mAP) and HACS Segments (32.09% mAP). Combined with an extra action classifier, it obtains 36.75% mAP on ActivityNet-1.3. Code is available at https://github.com/xlliu7/TadTR.

Current developments in temporal event or action localization usually target actions captured by a single camera. However, extensive events or actions in the wild may be captured as a sequence of shots by multiple cameras at different positions. In this paper, we propose a new and challenging task called multi-shot temporal event localization, and accordingly, collect a large-scale dataset called MUlti-Shot EventS (MUSES). MUSES has 31,477 event instances for a total of 716 video hours. The core nature of MUSES is the frequent shot cuts, for an average of 19 shots per instance and 176 shots per video, which induces large intra-instance variations. Our comprehensive evaluations show that the state-of-the-art method in temporal action localization only achieves an mAP of 13.1% at IoU=0.5. As a minor contribution, we present a simple baseline approach for handling the intra-instance variations, which reports an mAP of 18.9% on MUSES and 56.9% on THUMOS14 at IoU=0.5. To facilitate research in this direction, we release the dataset and the project code at https://songbai.site/muses/.

In this work we present SwiftNet for real-time semisupervised video object segmentation (one-shot VOS), which reports 77.8% $\mathcal{J}\& \mathcal{F}$ and 70 FPS on DAVIS 2017 validation dataset, leading all present solutions in overall accuracy and speed performance. We achieve this by elaborately compressing spatiotemporal redundancy in matching-based VOS via Pixel-Adaptive Memory (PAM). Temporally, PAM adaptively triggers memory updates on frames where objects display noteworthy inter-frame variations. Spatially, PAM selectively performs memory update and match on dynamic pixels while ignoring the static ones, significantly reducing redundant computations wasted on segmentation-irrelevant pixels. To promote efficient reference encoding, light-aggregation encoder is also introduced in SwiftNet deploying reversed sub-pixel. We hope SwiftNet could set a strong and efficient baseline for real-time VOS and facilitate its application in mobile vision. The source code of SwiftNet can be found at https://github.com/haochenheheda/SwiftNet.

Person search aims to simultaneously localize and identify a query person from realistic, uncropped images, which can be regarded as the unified task of pedestrian detection and person re-identification (re-id). Most existing works employ two-stage detectors like Faster-RCNN, yielding encouraging accuracy but with high computational overhead. In this work, we present the Feature-Aligned Person Search Network (AlignPS), the first anchor-free framework to efficiently tackle this challenging task. AlignPS explicitly addresses the major challenges, which we summarize as the misalignment issues in different levels (i.e., scale, region, and task), when accommodating an anchor-free detector for this task. More specifically, we propose an aligned feature aggregation module to generate more discriminative and robust feature embeddings by following a "re-id first" principle. Such a simple design directly improves the baseline anchor-free model on CUHK-SYSU by more than 20% in mAP. Moreover, AlignPS outperforms state-of-the-art two-stage methods, with a higher speed. The code is available at https://github.com/daodaofr/AlignPS.

This paper presents a neural network built upon Transformers, namely PlaneTR, to simultaneously detect and reconstruct planes from a single image. Different from previous methods, PlaneTR jointly leverages the context information and the geometric structures in a sequence-to-sequence way to holistically detect plane instances in one forward pass. Specifically, we represent the geometric structures as line segments and conduct the network with three main components: (i) context and line segments encoders, (ii) a structure-guided plane decoder, (iii) a pixel-wise plane embedding decoder. Given an image and its detected line segments, PlaneTR generates the context and line segment sequences via two specially designed encoders and then feeds them into a Transformers-based decoder to directly predict a sequence of plane instances by simultaneously considering the context and global structure cues. Finally, the pixel-wise embeddings are computed to assign each pixel to one predicted plane instance which is nearest to it in embedding space. Comprehensive experiments demonstrate that PlaneTR achieves state-of-the-art performance on the ScanNet and NYUv2 datasets.

In this paper, we propose a new task of deep interactive video inpainting and an application for users to interact with machines. To our best knowledge, this is the first deep learning-based interactive video inpainting framework that only uses a free form of user input as guidance (i.e. scribbles) instead of mask annotations, which has academic, entertainment, and commercial value.

Person re-identification (re-ID) has attracted much attention recently due to its great importance in video surveillance. In general, distance metrics used to identify two person images are expected to be robust under various appearance changes. However, our work observes the extreme vulnerability of existing distance metrics to adversarial examples, generated by simply adding human-imperceptible perturbations to person images. Hence, the security danger is dramatically increased when deploying commercial re-ID systems in video surveillance. Although adversarial examples have been extensively applied for classification analysis, it is rarely studied in metric analysis like person re-identification. The most likely reason is the natural gap between the training and testing of re-ID networks, that is, the predictions of a re-ID network cannot be directly used during testing without an effective metric. In this work, we bridge the gap by proposing Adversarial Metric Attack, a parallel methodology to adversarial classification attacks. Comprehensive experiments clearly reveal the adversarial effects in re-ID systems. Meanwhile, we also present an early attempt of training a metric-preserving network, thereby defending the metric against adversarial attacks. At last, by benchmarking various adversarial settings, we expect that our work can facilitate the development of adversarial attack and defense in metric-based applications.

This paper presents regional attraction of line segment maps, and hereby poses the problem of line segment detection (LSD) as a problem of region coloring. Given a line segment map, the proposed regional attraction first establishes the relationship between line segments and regions in the image lattice. Based on this, the line segment map is equivalently transformed to an attraction field map (AFM), which can be remapped to a set of line segments without loss of information. Accordingly, we develop an end-to-end framework to learn attraction field maps for raw input images, followed by a squeeze module to detect line segments. Apart from existing works, the proposed detector properly handles the local ambiguity and does not rely on the accurate identification of edge pixels. Comprehensive experiments on the Wireframe dataset and the YorkUrban dataset demonstrate the superiority of our method. In particular, we achieve an F-measure of 0.831 on the Wireframe dataset, advancing the state-of-the-art performance by 10.3 percent.

Recent development of adversarial attacks has proven that ensemble-based methods outperform traditional, non-ensemble ones in black-box attack. However, as it is computationally prohibitive to acquire a family of diverse models, these methods achieve inferior performance constrained by the limited number of models to be ensembled.In this paper, we propose Ghost Networks to improve the transferability of adversarial examples. The critical principle of ghost networks is to apply feature-level perturbations to an existing model to potentially create a huge set of diverse models. After that, models are subsequently fused by longitudinal ensemble. Extensive experimental results suggest that the number of networks is essential for improving the transferability of adversarial examples, but it is less necessary to independently train different networks and ensemble them in an intensive aggregation way. Instead, our work can be used as a computationally cheap and easily applied plug-in to improve adversarial approaches both in single-model and multi-model attack, compatible with residual and non-residual networks. By reproducing the NeurIPS 2017 adversarial competition, our method outperforms the No.1 attack submission by a large margin, demonstrating its effectiveness and efficiency. Code is available at https://github.com/LiYingwei/ghost-network.

Non-Local (NL) blocks have been widely studied in various vision tasks. However, it has been rarely explored to embed the NL blocks in mobile neural networks, mainly due to the following challenges: 1) NL blocks generally have heavy computation cost which makes it difficult to be applied in applications where computational resources are limited, and 2) it is an open problem to discover an optimal configuration to embed NL blocks into mobile neural networks. We propose AutoNL to overcome the above two obstacles. Firstly, we propose a Lightweight Non-Local (LightNL) block by squeezing the transformation operations and incorporating compact features. With the novel design choices, the proposed LightNL block is 400 times computationally cheaper} than its conventional counterpart without sacrificing the performance. Secondly, by relaxing the structure of the LightNL block to be differentiable during training, we propose an efficient neural architecture search algorithm to learn an optimal configuration of LightNL blocks in an end-to-end manner. Notably, using only 32 GPU hours, the searched AutoNL model achieves 77.7% top-1 accuracy on ImageNet under a typical mobile setting (350M FLOPs), significantly outperforming previous mobile models including MobileNetV2 (+5.7%), FBNet (+2.8%) and MnasNet (+2.1%). Code and models are available at https://github.com/LiYingwei/AutoNL.

This paper presents a fast and parsimonious parsing method to accurately and robustly detect a vectorized wireframe in an input image with a single forward pass. The proposed method is end-to-end trainable, consisting of three components: (i) line segment and junction proposal generation, (ii) line segment and junction matching, and (iii) line segment and junction verification. For computing line segment proposals, a novel exact dual representation is proposed which exploits a parsimonious geometric reparameterization for line segments and forms a holistic 4-dimensional attraction field map for an input image. Junctions can be treated as the “basins” in the attraction field. The proposed method is thus called Holistically-Attracted Wireframe Parser (HAWP). In experiments, the proposed method is tested on two benchmarks, the Wireframe dataset [14] and the YorkUrban dataset [8]. On both benchmarks, it obtains state-of-the-art performance in terms of accuracy and efficiency. For example, on the Wireframe dataset, compared to the previous state-of-the-art method L-CNN [36], it improves the challenging mean structural average precision (msAP) by a large margin (2.8% absolute improvements), and achieves 29.5 FPS on a single GPU (89% relative improvement). A systematic ablation study is performed to further justify the proposed method.

We propose a robust baseline method for instance segmentation which are specially designed for large-scale outdoor LiDAR point clouds. Our method includes a novel dense feature encoding technique, allowing the localization and segmentation of small, far-away objects, a simple but effective solution for single-shot instance prediction and effective strategies for handling severe class imbalances. Since there is no public dataset for the study of LiDAR instance segmentation, we also build a new publicly available LiDAR point cloud dataset to include both precise 3D bounding box and point-wise labels for instance segmentation, while still being about 3~20 times as large as other existing LiDAR datasets. The dataset will be published at https://github.com/feihuzhang/LiDARSeg.

In this paper, we focus on the semantic image synthesis task that aims at transferring semantic label maps to photo-realistic images. Existing methods lack effective semantic constraints to preserve the semantic information and ignore the structural correlations in both spatial and channel dimensions, leading to unsatisfactory blurry and artifact-prone results. To address these limitations, we propose a novel Dual Attention GAN (DAGAN) to synthesize photo-realistic and semantically-consistent images with fine details from the input layouts without imposing extra training overhead or modifying the network architectures of existing methods. We also propose two novel modules, i.e., position-wise Spatial Attention Module (SAM) and scale-wise Channel Attention Module (CAM), to capture semantic structure attention in spatial and channel dimensions, respectively. Specifically, SAM selectively correlates the pixels at each position by a spatial attention map, leading to pixels with the same semantic label being related to each other regardless of their spatial distances. Meanwhile, CAM selectively emphasizes the scale-wise features at each channel by a channel attention map, which integrates associated features among all channel maps regardless of their scales. We finally sum the outputs of SAM and CAM to further improve feature representation. Extensive experiments on four challenging datasets show that DAGAN achieves remarkably better results than state-of-the-art methods, while using fewer model parameters.

Weakly Supervised Object Detection (WSOD), using only image-level annotations to train object detectors, is of growing importance in object recognition. In this paper, we propose a novel deep network for WSOD. Unlike previous networks that transfer the object detection problem to an image classification problem using Multiple Instance Learning (MIL), our strategy generates proposal clusters to learn refined instance classifiers by an iterative process. The proposals in the same cluster are spatially adjacent and associated with the same object. This prevents the network from concentrating too much on parts of objects instead of whole objects. We first show that instances can be assigned object or background labels directly based on proposal clusters for instance classifier refinement, and then show that treating each cluster as a small new bag yields fewer ambiguities than the directly assigning label method. The iterative instance classifier refinement is implemented online using multiple streams in convolutional neural networks, where the first is an MIL network and the others are for instance classifier refinement supervised by the preceding one. Experiments are conducted on the PASCAL VOC, ImageNet detection, and MS-COCO benchmarks for WSOD. Results show that our method outperforms the previous state of the art significantly.

Learning robust and discriminative representations is essential for 3D object retrieval. In this paper, we present an improved Multi-view Convolutional Neural Network (MVCNN) for view-based 3D object representation learning. Our technical contributions are divided into two aspects. First, we propose to employ Group-view Similarity Learning (GSL) over the multi-view representations before the aggregation operation (i.e., max-pooling in MVCNN). We assume that the similarity information among the view groups of different 3D objects can provide an important cue but has been neglected more or less by previous methods. To enhance it, we add a branch to the original MVCNN architecture and learn to maintain such group-view similarity relationships. Second, we utilize an end-to-end metric learning loss function to improve the representation learning process. In particular, we propose an improved Triplet-Center Loss (TCL) named Adaptive Margin based Triplet-Center Loss (AMTCL). The original TCL assumes a fixed and common margin to control the relative distance relationship between a sample to its corresponding class center and to the nearest negative center. Though TCL has demonstrated its great capacity on the 3D object retrieval task, however, when considering the distinguishability between samples of one class and samples of another class, we assume that it would be more appropriate that the margin takes different values based on the distinguishability of samples of different classes. Therefore we propose to adaptively and dynamically adjust the margin hyperparameter based on the normalized confusion matrix which is obtained on the training set during the training process. Extensive experiments on several public 3D shape benchmarks show that our method, GSL + AMTCL, can learn more suitable representations for 3D object retrieval, obtaining superior performance against state-of-the-art methods.

This work studies the unsupervised re-ranking procedure for object retrieval and person re-identification with a specific concentration on an ensemble of multiple metrics (or similarities). While the re-ranking step is involved by running a diffusion process on the underlying data manifolds, the fusion step can leverage the complementarity of multiple metrics. We give a comprehensive summary of existing fusion with diffusion strategies, and systematically analyze their pros and cons. Based on the analysis, we propose a unified yet robust algorithm which inherits their advantages and discards their disadvantages. Hence, we call it Unified Ensemble Diffusion (UED). More interestingly, we derive that the inherited properties indeed stem from a theoretical framework, where the relevant works can be elegantly summarized as special cases of UED by imposing additional constraints on the objective function and varying the solver of similarity propagation. Extensive experiments with 3D shape retrieval, image retrieval and person re-identification demonstrate that the proposed framework outperforms the state of the arts, and at the same time suggest that re-ranking via metric fusion is a promising tool to further improve the retrieval performance of existing algorithms.

Though CNNs have achieved the state-of-the-art performance on various vision tasks, they are vulnerable to adversarial examples --- crafted by adding human-imperceptible perturbations to clean images. However, most of the existing adversarial attacks only achieve relatively low success rates under the challenging black-box setting, where the attackers have no knowledge of the model structure and parameters. To this end, we propose to improve the transferability of adversarial examples by creating diverse input patterns. Instead of only using the original images to generate adversarial examples, our method applies random transformations to the input images at each iteration. Extensive experiments on ImageNet show that the proposed attack method can generate adversarial examples that transfer much better to different networks than existing baselines. By evaluating our method against top defense solutions and official baselines from NIPS 2017 adversarial competition, the enhanced attack reaches an average success rate of 73.0%, which outperforms the top-1 attack submission in the NIPS competition by a large margin of 6.6%. We hope that our proposed attack strategy can serve as a strong benchmark baseline for evaluating the robustness of networks to adversaries and the effectiveness of different defense methods in the future. Code is available at https://github.com/cihangxie/DI-2-FGSM.

This paper presents a region-partition based attraction field dual representation for line segment maps, and thus poses the problem of line segment detection (LSD) as the region coloring problem. The latter is then addressed by learning deep convolutional neural networks (ConvNets) for accuracy, robustness and efficiency. For a 2D line segment map, our dual representation consists of three components: (i) A region-partition map in which every pixel is assigned to one and only one line segment; (ii) An attraction field map in which every pixel in a partition region is encoded by its 2D projection vector w.r.t. the associated line segment; and (iii) A squeeze module which squashes the attraction field to a line segment map that almost perfectly recovers the input one. By leveraging the duality, we learn ConvNets to compute the attraction field maps for raw in-put images, followed by the squeeze module for LSD, in an end-to-end manner. Our method rigorously addresses several challenges in LSD such as local ambiguity and class imbalance. Our method also harnesses the best practices developed in ConvNets based semantic segmentation methods such as the encoder-decoder architecture and the a-trous convolution. In experiments, our method is tested on the WireFrame dataset and the YorkUrban dataset with state-of-the-art performance obtained. Especially, we advance the performance by 4.5 percents on the WireFramedataset. Our method is also fast with 6.6∼10.4 FPS, outperforming most of existing line segment detectors.

In object detection, keypoint-based approaches often experience the drawback of a large number of incorrect object bounding boxes, arguably due to the lack of an additional assessment inside cropped regions. This paper presents an efficient solution that explores the visual patterns within individual cropped regions with minimal costs. We build our framework upon a representative one-stage keypoint-based detector named CornerNet. Our approach, named CenterNet, detects each object as a triplet, rather than a pair, of keypoints, which improves both precision and recall. Accordingly, we design two customized modules, cascade corner pooling, and center pooling, that enrich information collected by both the top-left and bottom-right corners and provide more recognizable information from the central regions. On the MS-COCO dataset, CenterNet achieves an AP of 47.0 %, outperforming all existing one-stage detectors by at least 4.9%. Furthermore, with a faster inference speed than the top-ranked two-stage detectors, CenterNet demonstrates a comparable performance to these detectors. Code is available at https://github.com/Duankaiwen/CenterNet.

How to aggregate multi-view representations of a 3D object into an informative and discriminative one remains a key challenge for multi-view 3D object retrieval. Existing methods either use view-wise pooling strategies which neglect the spatial information across different views or employ recurrent neural networks which may face the efficiency problem. To address these issues, we propose an effective and efficient framework called View N-gram Network (VNN). Inspired by n-gram models in natural language processing, VNN divides the view sequence into a set of visual n-grams, which involve overlapping consecutive view sub-sequences. By doing so, spatial information across multiple views is captured, which helps to learn a discriminative global embedding for each 3D object. Experiments on 3D shape retrieval benchmarks, including ModelNet10, ModelNet40 and ShapeNetCore55 datasets, demonstrate the superiority of our proposed method.

Dense crowd counting aims to predict thousands of human instances from an image, by calculating integrals of a density map over image pixels. Existing approaches mainly suffer from the extreme density variations. Such density pattern shift poses challenges even for multi-scale model ensembling. In this paper, we propose a simple yet effective approach to tackle this problem. First, a patch-level density map is extracted by a density estimation model and further grouped into several density levels which are determined over full datasets. Second, each patch density map is automatically normalized by an online center learning strategy with a multipolar center loss. Such a design can significantly condense the density distribution into several clusters, and enable that the density variance can be learned by a single model. Extensive experiments demonstrate the superiority of the proposed method. Our work outperforms the state-of-the-art by 4.2%, 14.3%, 27.1% and 20.1% in MAE, on the ShanghaiTech Part A, ShanghaiTech Part B, UCF_CC_50 and UCF-QNRF datasets, respectively.

Reading text in the wild is a very challenging task due to the diversity of text instances and the complexity of natural scenes. Recently, the community has paid increasing attention to the problem of recognizing text instances with irregular shapes. One intuitive and effective way to handle this problem is to rectify irregular text to a canonical form before recognition. However, these methods might struggle when dealing with highly curved or distorted text instances. To tackle this issue, we propose in this paper a Symmetry-constrained Rectification Network (ScRN) based on local attributes of text instances, such as center line, scale and orientation. Such constraints with an accurate description of text shape enable ScRN to generate better rectification results than existing methods and thus lead to higher recognition accuracy. Our method achieves state-of-the-art performance on text with both regular and irregular shapes. Specifically, the system outperforms existing algorithms by a large margin on datasets that contain quite a proportion of irregular text instances, e.g., ICDAR 2015, SVT-Perspective and CUTE80.

Unsupervised video object segmentation has often been tackled by methods based on recurrent neural networks and optical flow. Despite their complexity, these kinds of approach tend to favour short-term temporal dependencies and are thus prone to accumulating inaccuracies, which cause drift over time. Moreover, simple (static) image segmentation models, alone, can perform competitively against these methods, which further suggests that the way temporal dependencies are modelled should be reconsidered. Motivated by these observations, in this paper we explore simple yet effective strategies to model long-term temporal dependencies. Inspired by the non-local operators, we introduce a technique to establish dense correspondences between pixel embeddings of a reference "anchor" frame and the current one. This allows the learning of pairwise dependencies at arbitrarily long distances without conditioning on intermediate frames. Without online supervision, our approach can suppress the background and precisely segment the foreground object even in challenging scenarios, while maintaining consistent performance over time. With a mean IoU of 81.7%, our method ranks first on the DAVIS-2016 leaderboard of unsupervised methods, while still being competitive against state-of-the-art online semi-supervised approaches. We further evaluate our method on the FBMS dataset and the video saliency dataset ViSal, showing results competitive with the state of the art.

Accurate multi-organ abdominal CT segmentation is essential to many clinical applications such as computer-aided intervention. As data annotation requires massive human labor from experienced radiologists, it is common that training data is usually partially-labeled. However, these background labels can be misleading in multi-organ segmentation since the ``background'' usually contains some other organs of interest. To address the background ambiguity in these partially-labeled datasets, we propose Prior-aware Neural Network (PaNN) via explicitly incorporating anatomical priors on abdominal organ sizes, guiding the training process with domain-specific knowledge. More specifically, PaNN assumes that the average organ size distributions in the abdomen should approximate their empirical distributions, a prior statistics obtained from the fully-labeled dataset. As our objective is difficult to be directly optimized using stochastic gradient descent, it is reformulated as a min-max form and optimized via the stochastic primal-dual gradient algorithm. PaNN achieves state-of-the-art performance on the MICCAI2015 challenge ``Multi-Atlas Labeling Beyond the Cranial Vault'', a competition on organ segmentation in the abdomen. We report an average Dice score of 84.97%, surpassing the prior art by a large margin of 3.27%. Code and models will be made publicly available.

The non-local module works as a particularly useful technique for semantic segmentation while criticized for its prohibitive computation and GPU memory occupation. In this paper, we present Asymmetric Non-local Neural Network to semantic segmentation, which has two prominent components: Asymmetric Pyramid Non-local Block (APNB) and Asymmetric Fusion Non-local Block (AFNB). APNB leverages a pyramid sampling module into the non-local block to largely reduce the computation and memory consumption without sacrificing the performance. AFNB is adapted from APNB to fuse the features of different levels under a sufficient consideration of long range dependencies and thus considerably improves the performance. Extensive experiments on semantic segmentation benchmarks demonstrate the effectiveness and efficiency of our work. In particular, we report the state-of-the-art performance of 81.3 mIoU on the Cityscapes test set. For a 256x128 input, APNB is around 6 times faster than a non-local block on GPU while 28 times smaller in GPU running memory occupation. Code is available at: https://github.com/MendelXu/ANN.git.

In multi-organ segmentation of abdominal CT scans, most existing fully supervised deep learning algorithms require lots of voxel-wise annotations, which are usually difficult, expensive, and slow to obtain. In comparison, massive unlabeled 3D CT volumes are usually easily accessible. Current mainstream works to address semi-supervised biomedical image segmentation problem are mostly graph-based. By contrast, deep network based semi-supervised learning methods have not drawn much attention in this field. In this work, we propose Deep Multi-Planar Co-Training (DMPCT), whose contributions can be divided into two folds: 1) The deep model is learned in a co-training style which can mine consensus information from multiple planes like the sagittal, coronal, and axial planes; 2) Multi-planar fusion is applied to generate more reliable pseudo-labels, which alleviates the errors occurring in the pseudo-labels and thus can help to train better segmentation networks. Experiments are done on our newly collected large dataset with 100 unlabeled cases as well as 210 labeled cases where 16 anatomical structures are manually annotated by four radiologists and confirmed by a senior expert. The results suggest that DMPCT significantly outperforms the fully supervised method by more than 4% especially when only a small set of annotations is used.

Diffusion process has advanced object retrieval greatly as it can capture the underlying manifold structure. Recent studies have experimentally demonstrated that tensor product diffusion can better reveal the intrinsic relationship between objects than other variants. However, the principle remains unclear, i.e., what kind of manifold structure is captured. In this paper, we propose a new affinity learning algorithm called Regularized Diffusion Process (RDP). By deeply exploring the properties of RDP, our first yet basic contribution is providing a manifold-based explanation for tensor product diffusion. A novel criterion measuring the smoothness of the manifold is defined, which simultaneously regularizes four vertices in the affinity graph. Inspired by this observation, we further contribute two variants towards two specific goals. While ARDP can learn similarities across heterogeneous domains, HRDP performs affinity learning on tensor product hypergraph, considering the relationships between objects are generally more complex than pairwise. Consequently, RDP, ARDP and HRDP constitute a generic tool for object retrieval in most commonly-used settings, no matter the input relationships between objects are derived from the same domain or not, and in pairwise formulation or not. Comprehensive experiments on 10 retrieval benchmarks, especially on large scale data, validate the effectiveness and generalization of our work.

As a post-processing procedure, the diffusion process has demonstrated its ability of substantially improving the performance of various visual retrieval systems. Whereas, great efforts are also devoted to similarity (or metric) fusion, seeing that only one individual type of similarity cannot fully reveal the intrinsic relationship between objects. This stimulates a great research interest of considering similarity fusion in the framework of the diffusion process (i.e., fusion with diffusion) for robust retrieval. In this paper, we first revisit representative methods about fusion with diffusion and provide new insights which are ignored by previous researchers. Then, observing that existing algorithms are susceptible to noisy similarities, the proposed regularized ensemble diffusion (RED) is bundled with an automatic weight learning paradigm, so that the negative impacts of noisy similarities are suppressed. Though formulated as a convex optimization problem, one advantage of RED is that it converts back into the iteration-based solver with the same computational complexity as the conventional diffusion process. At last, we integrate several recently-proposed similarities with the proposed framework. The experimental results suggest that we can achieve new state-of-the-art performances on various retrieval tasks, including 3D shape retrieval on the ModelNet data set, and image retrieval on the Holidays and Ukbench data sets.

Most existing 3D object recognition algorithms focus on leveraging the strong discriminative power of deep learning models with softmax loss for the classification of 3D data, while learning discriminative features with deep metric learning for 3D object retrieval is more or less neglected. In the paper, we study variants of deep metric learning losses for 3D object retrieval, which did not receive enough attention from this area. First, two kinds of representative losses, triplet loss and center loss, are introduced which could learn more discriminative features than traditional classification loss. Then, we propose a novel loss named triplet-center loss, which can further enhance the discriminative power of the features. The proposed triplet-center loss learns a center for each class and requires that the distances between samples and centers from the same class are closer than those from different classes. Extensive experimental results on two popular 3D object retrieval benchmarks and two widely-adopted sketch-based 3D shape retrieval benchmarks consistently demonstrate the effectiveness of our proposed loss, and significant improvements have been achieved compared with the state-of-the-arts.

Multidimensional Scaling (MDS) is a classic technique that seeks vectorial representations for data points, given the pairwise distances between them. In recent years, data are usually collected from diverse sources or have multiple heterogeneous representations. However, how to do multidimensional scaling on multiple input distance matrices is still unsolved to our best knowledge. In this paper, we first define this new task formally. Then, we propose a new algorithm called Multi-View Multidimensional Scaling (MVMDS) by considering each input distance matrix as one view. The proposed algorithm can learn the weights of views (i.e., distance matrices) automatically by exploring the consensus information and complementary nature of views. Experimental results on synthetic as well as real datasets demonstrate the effectiveness of MVMDS. We hope that our work encourages a wider consideration in many domains where MDS is needed.

Diffusion process has advanced visual retrieval greatly owing to its capacity in capturing the geometry structure of the underlying manifold. Recent studies (Donoser and Bischof 2013) have experimentally demonstrated that diffusion process on the tensor product graph yields better retrieval performances than that on the original affinity graph. However, the principle behind this kind of diffusion process remains unclear, i.e., what kind of manifold structure is captured and how it is reflected. In this paper, we propose a new variant o diffusion process, which also operates on a tensor product graph. It is defined in three equivalent formulations (regularization framework, iterative framework and limit framework, respectively). Based on our study, three insightful conclusions are drawn which theoretically explain how this kind of diffusion process can better reveal the intrinsic relationship between objects. Besides, extensive experimental results on various retrieval tasks testify the validity of the proposed method.

Most existing person re-identification algorithms either extract robust visual features or learn discriminative metrics for person images. However, the underlying manifold which those images reside on is rarely investigated. That arises a problem that the learned metric is not smooth with respect to the local geometry structure of the data manifold. In this paper, we study person re-identification with manifold-based affinity learning, which did not receive enough attention from this area. An unconventional manifold-preserving algorithm is proposed, which can 1) make best use of supervision from training data, whose label information is given as pairwise constraints, 2) scale up to large repositories with low on-line time complexity, and 3) be plunged into most existing algorithms, serving as a generic postprocessing procedure to further boost the identification accuracies. Extensive experimental results on five popular person re-identification benchmarks consistently demonstrate the effectiveness of our method. Especially, on the largest CUHK03 and Market-1501, our method outperforms the state-of-the-art alternatives by a large margin with high efficiency, which is more appropriate for practical applications.

As a postprocessing procedure, diffusion process has demonstrated its ability of substantially improving the performance of various visual retrieval systems. Whereas, great efforts are also devoted to similarity (or metric) fusion, seeing that only one individual type of similarity cannot fully reveal the intrinsic relationship between objects. This stimulates a great research interest of considering similarity fusion in the framework of diffusion process (i.e., fusion with diffusion) for robust retrieval. In this paper, we firstly revisit representative methods about fusion with diffusion, and provide new insights which are ignored by previous researchers. Then, observing that existing algorithms are susceptible to noisy similarities, the proposed Regularized Ensemble Diffusion (RED) is bundled with an automatic weight learning paradigm, so that the negative impacts of noisy similarities are suppressed. At last, we integrate several recently-proposed similarities with the proposed framework. The experimental results suggest that we can achieve new state-of-the-art performances on various retrieval tasks, including 3D shape retrieval on ModelNet dataset, and image retrieval on Holidays and Ukbench dataset.

Projective analysis is an important solution in three-dimensional (3D) shape retrieval, since human visual perceptions of 3D shapes rely on various 2D observations from different viewpoints. Although multiple informative and discriminative views are utilized, most projection-based retrieval systems suffer from heavy computational cost, and thus cannot satisfy the basic requirement of scalability for search engines. In the past three years, shape retrieval contest (SHREC) pays much attention to the scalability of 3D shape retrieval algorithms, and organizes several large scale tracks accordingly [1]- [3]. However, the experimental results indicate that conventional algorithms cannot be directly applied to large datasets. In this paper, we present a real-time 3D shape search engine based on the projective images of 3D shapes. The real-time property of our search engine results from the following aspects: (1) efficient projection and view feature extraction using GPU acceleration; (2) the first inverted file, called F-IF, is utilized to speed up the procedure of multiview matching; and (3) the second inverted file, which captures a local distribution of 3D shapes in the feature manifold, is adopted for efficient context-based reranking. As a result, for each query the retrieval task can be finished within one second despite the necessary cost of IO overhead. We name the proposed 3D shape search engine, which combines GPU acceleration and inverted file (t wice), as GIFT. Besides its high efficiency, GIFT also outperforms state-of-the-art methods significantly in retrieval accuracy on various shape benchmarks (ModelNet40 dataset, ModelNet10 dataset, PSB dataset, McGill dataset) and competitions (SHREC14LSGTB, ShapeNet Core55, WM-SHREC07).

Projective analysis is an important solution for 3D shape retrieval, since human visual perceptions of 3D shapes rely on various 2D observations from different view points. Although multiple informative and discriminative views are utilized, most projection-based retrieval systems suffer from heavy computational cost, thus cannot satisfy the basic requirement of scalability for search engines. In this paper, we present a real-time 3D shape search engine based on the projective images of 3D shapes. The real-time property of our search engine results from the following aspects: (1) efficient projection and view feature extraction using GPU acceleration, (2) the first inverted file, referred as F-IF, is utilized to speed up the procedure of multi-view matching, (3) the second inverted file (S-IF), which captures a local distribution of 3D shapes in the feature manifold, is adopted for efficient context-based reranking. As a result, for each query the retrieval task can be finished within one second despite the necessary cost of IO overhead. We name the proposed 3D shape search engine, which combines GPU acceleration and Inverted File (Twice), as GIFT. Besides its high efficiency, GIFT also outperforms the state-of-the-art methods significantly in retrieval accuracy on various shape benchmarks and competitions.

In this paper, we propose an extremely efficient algorithm for visual re-ranking. By considering the original pairwise distance in the contextual space, we develop a feature vector called sparse contextual activation (SCA) that encodes the local distribution of an image. Hence, re-ranking task can be simply accomplished by vector comparison under the generalized Jaccard metric, which has its theoretical meaning in the fuzzy set theory. In order to improve the time efficiency of re-ranking procedure, inverted index is successfully introduced to speed up the computation of generalized Jaccard metric. As a result, the average time cost of re-ranking for a certain query can be controlled within 1 ms. Furthermore, inspired by query expansion, we also develop an additional method called local consistency enhancement on the proposed SCA to improve the retrieval performance in an unsupervised manner. On the other hand, the retrieval performance using a single feature may not be satisfactory enough, which inspires us to fuse multiple complementary features for accurate retrieval. Based on SCA, a robust feature fusion algorithm is exploited that also preserves the characteristic of high time efficiency. We assess our proposed method in various visual re-ranking tasks. Experimental results on Princeton shape benchmark (3D object), WM-SRHEC07 (3D competition), YAEL data set B (face), MPEG-7 data set (shape), and Ukbench data set (image) manifest the effectiveness and efficiency of SCA.

Feature coding is a fundamental issue with many vision tasks, such as image classification, image retrieval and image segmentation, etc. There is no doubt that the encoding procedure leads to information loss, due to the existence of quantization error. The residual vector, defined as the difference between the feature and its corresponding visual word, is the chief culprit to be responsible for the quantization error. Many previous algorithms consider it as a coding issue, and focus on reducing the quantization error by reconstructing the feature with more than one visual word, or by the so-called soft-assignment strategy. In this paper, we consider the problem from a different point of view, and propose an effective and efficient model called multiple stage residual model (MSRM). It makes full use of the residual vector to generate a multiple stage code. MSRM is a hierarchical structure, with the bottom stage producing the coarsest quantization, and the top stage producing the finest quantization. Moreover, our proposed model is a generic framework, which can be built upon many coding algorithms. The interplay of such a coarse-to-fine quantization procedure with a discriminative classifier (e.g., SVM) can improve the classification accuracy of the baseline algorithms significantly. As a special case of MSRM, multiple stage vector quantization (MSVQ) can be directly used for vector compression and approximate nearest neighbor search, and achieves competitive performances with high efficiency.

View-based 3D shape retrieval is a popular branch in 3D shape analysis owing to the high discriminative property of 2D views. However, many previous works do not scale up to large 3D shape databases. We propose a two layer coding (TLC) framework to conduct shape matching much more efficiently. The first layer coding is applied to pairs of views represented as depth images. The spatial relationship of each view pair is captured with so-called eigen-angle, which is the planar angle between the two views measured at the center of the 3D shape. Prior to the second layer coding, the view pairs are divided into subsets according to their eigen-angles. Consequently, view pairs that differ significantly in their eigen-angles are encoded with different codewords, which implies that spatial arrangement of views is preserved in the second layer coding. The final feature vector of a 3D shape is the concatenation of all the encoded features from different subsets, which is used for efficient indexing directly. TLC is not limited to encode the local features from 2D views, but can be also applied to encoding 3D features. Exhaustive experimental results confirm that TLC achieves state-of-the-art performance in both retrieval accuracy and efficiency.

This letter introduces a robust representation of 3-D shapes, named DeepPano, learned with deep convolutional neural networks (CNN). Firstly, each 3-D shape is converted into a panoramic view, namely a cylinder projection around its principle axis. Then, a variant of CNN is specifically designed for learning the deep representations directly from such views. Different from typical CNN, a row-wise max-pooling layer is inserted between the convolution and fully-connected layers, making the learned representations invariant to the rotation around a principle axis. Our approach achieves state-of-the-art retrieval/classification results on two large-scale 3-D model datasets (ModelNet-10 and ModelNet-40), outperforming typical methods by a large margin.

In this paper, we address the problem of building a compact representation for shape. We first decompose shape into meaningful contour fragments, and each fragment is described by a certain descriptor, e.g., Shape Context. Then inspired by the coding scheme Vector of Locally Aggregated Descriptors widely used in image representation, we try to aggregate the contour fragments into a very compact vector of limited dimension to stand for a shape, and we name the new designed shape descriptor as Vector of Aggregated Contour Fragments (VACF). We apply VACF to shape classification task on the well-known MPEG-7 shape benchmark, and the experimental results show that the accuracy of our proposed method outperforms other state-of-the-art algorithms with much smaller memory usage.