Estimating the geometric and volumetric properties of transparent deformable liquids is challenging due to optical complexities and dynamic surface deformations induced by container movements. Autonomous robots performing precise liquid manipulation tasks—such as dispensing, aspiration, and mixing—must handle containers in ways that inevitably induce these deformations, complicating accurate liquid state assessment. Current datasets lack comprehensive physics-informed simulation data representing realistic liquid behaviors under diverse dynamic scenarios. To bridge this gap, we introduce Phys-Liquid, a physics-informed dataset comprising 97,200 simulation images and corresponding 3D meshes, capturing liquid dynamics across multiple laboratory scenes, lighting conditions, liquid colors, and container rotations. To validate the realism and effectiveness of Phys-Liquid, we propose a four-stage reconstruction and estimation pipeline involving liquid segmentation, multi-view mask generation, 3D mesh reconstruction, and real-world scaling. Experimental results demonstrate improved accuracy and consistency in reconstructing liquid geometry and volume, outperforming existing benchmarks. The dataset and associated validation methods facilitate future advancements in transparent liquid perception tasks.

Gait recognition offers a non-intrusive biometric solution by identifying individuals through their walking patterns. Although discriminative models have achieved notable success in this domain, the full potential of generative models remains largely unexplored. In this paper, we introduce CoD², a novel framework that combines the data distribution modeling capabilities of diffusion models with the semantic representation learning strengths of discriminative models to extract robust gait features. We propose a Multi-level Conditional Control strategy that integrates both high-level identity-aware semantic conditions and low-level visual details. Specifically, the high-level condition, extracted by the discriminative extractor, guides the generation of identity-consistent gait sequences, while low-level visual details, such as appearance and motion, are preserved to enhance consistency. Moreover, the generated sequences facilitate the discriminative extractor's learning, enabling it to capture more comprehensive high-level semantic features. Extensive experiments on four datasets (SUSTech1K, CCPG, GREW, and Gait3D) demonstrate that CoD² achieves state-of-the-art performance and can be seamlessly integrated with existing discriminative methods, yielding consistent improvements.

Optical Chemical Structure Recognition (OCSR) plays a pivotal role in modern chemical informatics, enabling the automated conversion of chemical structure images from scientific literature, patents, and educational materials into machine-readable molecular representations. This capability is essential for large-scale chemical data mining, drug discovery pipelines, and Large Language Model (LLM) applications in related domains. However, existing OCSR systems face significant challenges in accurately recognizing stereochemical information due to the subtle visual cues that distinguish stereoisomers, such as wedge and dash bonds, ring conformations, and spatial arrangements. To address these challenges, we propose MolSight, a comprehensive learning framework for OCSR that employs a three-stage training paradigm. In the first stage, we conduct pre-training on large-scale but noisy datasets to endow the model with fundamental perception capabilities for chemical structure images. In the second stage, we perform multi-granularity fine-tuning using datasets with richer supervisory signals, systematically exploring how auxiliary tasks—specifically chemical bond classification and atom localization—contribute to molecular formula recognition. Finally, we employ reinforcement learning for post-training optimization and introduce a novel stereochemical structure dataset. Remarkably, we find that even with MolSight's relatively compact parameter size, the Group Relative Policy Optimization (GRPO) algorithm can further enhance the model's performance on stereomolecular. Through extensive experiments across diverse datasets, our results demonstrate that MolSight achieves state-of-the-art performance in (stereo)chemical optical structure recognition.

Flow-based 3D generation models typically require dozens of sampling steps during inference. Though few-step distillation methods, particularly Consistency Models (CMs), have achieved substantial advancements in accelerating 2D diffusion models, they remain under-explored for more complex 3D generation tasks. In this study, we propose a novel framework, MDT-dist, for few-step 3D flow distillation. Our approach is built upon a primary objective: distilling the pretrained model to learn the Marginal-Data Transport. Directly learning this objective needs to integrate the velocity fields, while this integral is intractable to be implemented. Therefore, we propose two optimizable objectives, Velocity Matching (VM) and Velocity Distillation (VD), to equivalently convert the optimization target from the transport level to the velocity and the distribution level respectively. Velocity Matching (VM) learns to stably match the velocity fields between the student and the teacher, but inevitably provides biased gradient estimates. Velocity Distillation (VD) further enhances the optimization process by leveraging the learned velocity fields to perform probability density distillation. When evaluated on the pioneer 3D generation framework TRELLIS, our method reduces sampling steps of each flow transformer from 25 to 1–2, achieving 0.68s (1 step x2) and 0.94s (2 steps x2) latency with 9.0x and 6.5x speedup on A800, while preserving high visual and geometric fidelity. Experiments demonstrate that our method significantly outperforms existing CM distillation methods, and enables TRELLIS to achieve superior performance in few-step 3D generation.

Text-prompted image segmentation enables fine-grained visual understanding and is critical for applications such as human-computer interaction and robotics. However, existing supervised fine-tuning methods typically ignore explicit chain-of-thought (CoT) reasoning at test time, which limits their ability to generalize to unseen prompts and domains. To address this issue, we introduce LENS, a scalable reinforcement-learning framework that jointly optimizes the reasoning process and segmentation in an end-to-end manner. We propose unified reinforcement-learning rewards that span sentence-, box-, and segment-level cues, encouraging the model to generate informative CoT rationales while refining mask quality. Using a publicly available 3-billion-parameter vision–language model, i.e., Qwen2.5-VL-3B-Instruct, LENS achieves an average cIoU of 81.2% on the RefCOCO, RefCOCO+, and RefCOCOg benchmarks, outperforming the strong fine-tuned method, i.e., GLaMM, by up to 5.6%. These results demonstrate that RL-driven CoT reasoning significantly enhances text-prompted segmentation and offers a practical path toward more generalizable Segment Anything models (SAM).

There is a growing demand for deploying large generative AI models on mobile devices. For recent popular video generative models, however, the Variational AutoEncoder (VAE) represents one of the major computational bottlenecks. Both large parameter sizes and mismatched kernels cause out-of-memory errors or extremely slow inference on mobile devices. To address this, we propose a low-cost solution that efficiently transfers widely used video VAEs to mobile devices. (1) We analyze redundancy in existing VAE architectures and get empirical design insights. By integrating 3D depthwise separable convolutions into our model, we significantly reduce the number of parameters. (2) We observe that the upsampling techniques in mainstream video VAEs are poorly suited to mobile hardware and form the main bottleneck. In response, we propose a decoupled 3D pixel shuffle scheme that slashes end-to-end delay. Building upon these, we develop a universal mobile-oriented VAE decoder, Turbo-VAED. (3) We propose an efficient VAE decoder training method. Since only the decoder is used during deployment, we distill it to Turbo-VAED instead of retraining the full VAE, enabling fast mobile adaptation with minimal performance loss. To our knowledge, our method enables real-time 720p video VAE decoding on mobile devices for the first time. This approach is widely applicable to most video VAEs. When integrated into four representative models, with training cost as low as $95, it accelerates original VAEs by up to 84.5× at 720p resolution on GPUs, uses as low as 17.5% of original parameter count, and retains 96.9% of the original reconstruction quality. Compared to mobile-optimized VAEs, Turbo-VAED achieves a 2.9× speedup in FPS and better reconstruction quality on the iPhone 16 Pro.

Soft Facial Attribute Recognition (FAR) remains largely unexplored in terms of demographic fairness. To the best of our knowledge, this study presents one of the first comprehensive analyses of demographic bias in FAR, proposing a systematic framework to detect, quantify, and promote awareness of both representational and stereotypical biases, supporting their mitigation. Leveraging established taxonomies, we evaluate state-of-the-art datasets using a rigorous set of interpretable bias metrics to uncover hidden demographic imbalances. To support reliable fairness assessment, we first enrich the datasets with standardized demographic annotations using the FairFace model. We then address label inconsistencies through the integration of predictions from advanced Vision-Language Models (VLMs). Our analysis reveals substantial imbalances across gender, age, and racial categories-specifically White, Black, and Asian- affecting dataset composition. Furthermore, we show that conventional fairness metrics often yield divergent assessments, highlighting the importance of multi-metric evaluation. This study provides a replicable methodology and actionable insights to support bias-aware facial analysis.

Generating visual text in natural scene images is a challenging task with many unsolved problems. Different from generating text on artificially designed images (such as posters, covers, and cartoons), existing methods for natural scene visual text generation still have significant deficiencies: methods based on rendering engines rely on manually crafted rules, which struggle to adapt to diverse backgrounds and leave obvious artificial traces, while their text layouts may be placed in unreasonable areas (e.g., sky or ground) and text content is semantically disconnected from the scene; diffusion model-based methods, on the other hand, face difficulties in generating small characters, depend on manually designed prompts to ensure reasonable layout and content, fail to generate text at precise locations, and cannot effectively control text attributes (e.g., font and color). In this paper, we propose a two-stage method named SceneVTG++ to address these issues. SceneVTG++ comprises two core components: a Text Layout and Content Generator (TLCG) and a Controllable Local Text Diffusion (CLTD). The former leverages the world knowledge and visual reasoning capabilities of multimodal large language models to identify reasonable text areas and recommend scene-relevant text content based on natural scene background images; the latter generates controllable multilingual text using a diffusion model, ensuring alignment with the outputs of TLCG. Through extensive experiments, we verified the effectiveness of both TLCG and CLTD, and demonstrated that SceneVTG++ achieves state-of-the-art performance in natural scene visual text generation. Additionally, the images generated by SceneVTG++ exhibit superior utility for training natural scene optical character recognition (OCR) tasks, including text detection and text recognition. Codes and datasets will be made publicly available.

Transformer has been very successful in various computer vision tasks and understanding the working mechanism of transformer is important. As touchstones, weakly-supervised semantic segmentation (WSSS) and class activation map (CAM) are useful tasks for analyzing vision transformers (ViT). Based on the plain ViT pre-trained with ImageNet classification, we find that multi-layer, multi-head self-attention maps can provide rich and diverse information for weakly-supervised semantic segmentation and CAM generation, e.g., different attention heads of ViT focus on different image areas and object categories. Thus we propose a novel method to end-to-end estimate the importance of attention heads, where the self-attention maps are adaptively fused for high-quality CAM results that tend to have more complete objects. Besides, we propose a ViT-based gradient clipping decoder for online retraining with the CAM results efficiently and effectively. Furthermore, the gradient clipping decoder can make good use of the knowledge in large-scale pre-trained ViT and has a scalable ability. The proposed plain Transformer-based Weakly-supervised learning method (WeakTr) obtains the superior WSSS performance on standard benchmarks, i.e., 78.5% mIoU on the $val$ set of PASCAL VOC 2012 and 51.1% mIoU on the $val$ set of COCO 2014. Source code and checkpoints are available at https://github.com/hustvl/WeakTr.

Recently, linear complexity sequence modeling networks have achieved modeling capabilities similar to Vision Transformers on a variety of computer vision tasks, while using fewer FLOPs and less memory. However, their advantage in terms of actual runtime speed is not significant. To address this issue, we introduce Gated Linear Attention (GLA) for vision, leveraging its superior hardware-awareness and efficiency. We propose direction-wise gating to capture 1D global context through bidirectional modeling and a 2D gating locality injection to adaptively inject 2D local details into 1D global context. Our hardware-aware implementation further merges forward and backward scanning into a single kernel, enhancing parallelism and reducing memory cost and latency. The proposed model, ViG, offers a favorable trade-off in accuracy, parameters, and FLOPs on ImageNet and downstream tasks, outperforming popular Transformer and CNN-based models.

The rapid and accurate direct multi-frame interpolation method for Digital Subtraction Angiography (DSA) images is crucial for reducing radiation and providing real-time assistance to physicians for precise diagnostics and treatment. DSA images contain complex vascular structures and various motions. Applying natural scene Video Frame Interpolation (VFI) methods results in motion artifacts, structural dissipation, and blurriness. Recently, MoSt-DSA has specifically addressed these issues for the first time and achieved SOTA results. However, MoSt-DSA's focus on real-time performance leads to insufficient suppression of high-frequency noise and incomplete filtering of low-frequency noise in the generated images. To address these issues within the same computational time scale, we propose GaraMoSt. Specifically, we optimize the network pipeline with a parallel design and propose a module named MG-MSFE. MG-MSFE extracts frame-relative motion and structural features at various granularities in a fully convolutional parallel manner and supports independent, flexible adjustment of context-aware granularity at different scales, thus enhancing computational efficiency and accuracy. Extensive experiments demonstrate that GaraMoSt achieves the SOTA performance in accuracy, robustness, visual effects, and noise suppression, comprehensively surpassing MoSt-DSA and other natural scene VFI methods.

3D Semantic Occupancy Prediction is fundamental for spatial understanding, yet existing approaches face challenges in scalability and generalization due to their reliance on extensive labeled data and computationally intensive voxel-wise representations. In this paper, we introduce GaussTR, a novel Gaussian-based TRansformer framework that unifies sparse 3D modeling with foundation model alignment through Gaussian representations to advance 3D spatial understanding. GaussTR predicts sparse sets of Gaussians in a feed-forward manner to represent 3D scenes. By splatting the Gaussians into 2D views and aligning the rendered features with foundation models, GaussTR facilitates self-supervised 3D representation learning and enables open-vocabulary semantic occupancy prediction without requiring explicit annotations. Empirical experiments on the Occ3D-nuScenes dataset demonstrate GaussTR’s state-of-the-art zero-shot performance of 12.27 mIoU, along with a 40% reduction in training time. These results highlight the efficacy of GaussTR for scalable and holistic 3D spatial understanding, with promising implications in autonomous driving and embodied agents. The code is available at https://github.com/hustvl/GaussTR.

Recent open-vocabulary segmentation methods adopt mask generators to predict segmentation masks and leverage pretrained vision-language models, e.g., CLIP, to classify these masks via mask pooling. Although these approaches show promising results, it is counterintuitive that accurate masks often fail to yield accurate classification results through pooling CLIP image embeddings within the mask regions. In this paper, we reveal the performance limitations of mask pooling and introduce Mask-Adapter, a simple yet effective method to address these challenges in open-vocabulary segmentation. Compared to directly using proposal masks, our proposed Mask-Adapter extracts semantic activation maps from proposal masks, providing richer contextual information and ensuring alignment between masks and CLIP. Additionally, we propose a mask consistency loss that encourages proposal masks with similar IoUs to obtain similar CLIP embeddings to enhance models’ robustness to varying predicted masks. Mask-Adapter integrates seamlessly into open-vocabulary segmentation methods based on mask pooling in a plug-and-play manner, delivering more accurate classification results. Extensive experiments across several zero-shot benchmarks demonstrate significant performance gains for the proposed Mask-Adapter on several well-established methods. Notably, Mask-Adapter also extends effectively to SAM and achieves impressive results on several open-vocabulary segmentation datasets. Code and models are available at https://github.com/hustvl/MaskAdapter.

Recently, the diffusion model has emerged as a powerful generative technique for robotic policy learning, capable of modeling multi-mode action distributions. Leveraging its capability for end-to-end autonomous driving is a promising direction. However, the numerous denoising steps in the robotic diffusion policy and the more dynamic, open-world nature of traffic scenes pose substantial challenges for generating diverse driving actions at a real-time speed. To address these challenges, we propose a novel truncated diffusion policy that incorporates prior multi-mode anchors and truncates the diffusion schedule, enabling the model to learn denoising from anchored Gaussian distribution to the multi-mode driving action distribution. Additionally, we design an efficient cascade diffusion decoder for enhanced interaction with conditional scene context. The proposed model, DiffusionDrive, demonstrates 10× reduction in denoising steps compared to vanilla diffusion policy, delivering superior diversity and quality in just 2 steps. On the planning-oriented NAVSIM dataset, with aligned ResNet-34 backbone, DiffusionDrive achieves 88.1 PDMS without bells and whistles, setting a new record, while running at a real-time speed of 45 FPS on an NVIDIA 4090. Qualitative results on challenging scenarios further confirm that DiffusionDrive can robustly generate diverse plausible driving actions.

Latent diffusion models with Transformer architectures excel at generating high-fidelity images. However, recent studies reveal an optimization dilemma in this two-stage design: while increasing the per-token feature dimension in visual tokenizers improves reconstruction quality, it requires substantially larger diffusion models and more training iterations to achieve comparable generation performance. Consequently, existing systems often settle for suboptimal solutions, either producing visual artifacts due to information loss within tokenizers or failing to converge fully due to expensive computation costs. We argue that this dilemma stems from the inherent difficulty in learning unconstrained high-dimensional latent spaces. To address this, we propose aligning the latent space with pretrained vision foundation models when training the visual tokenizers. Our proposed VA-VAE (Vision foundation model Aligned Variational AutoEncoder) significantly expands the reconstruction-generation frontier of latent diffusion models, enabling faster convergence of Diffusion Transformers (DiT) in high-dimensional latent spaces. To exploit the full potential of VA-VAE, we build an enhanced DiT baseline with improved training strategies and architecture designs, termed LightningDiT. The integrated system achieves state-of-the-art (SOTA) performance on ImageNet 256×256 generation with an FID score of 1.35 while demonstrating remarkable training efficiency by reaching an FID score of 2.11 in just 64 epochs – representing an over 21× convergence speedup compared to the original DiT. Models and codes are available at https://github.com/hustvl/LightningDiT.

Diffusion models with large-scale pre-training have achieved significant success in the field of visual content generation, particularly exemplified by Diffusion Transformers (DiT). However, DiT models have faced challenges with quadratic complexity efficiency, especially when handling long sequences. In this paper, we aim to incorporate the sub-quadratic modeling capability of Gated Linear Attention (GLA) into the 2D diffusion backbone. Specifically, we introduce Diffusion Gated Linear Attention Transformers (DiG), a simple, adoptable solution with minimal parameter overhead. We offer two variants, i,e, a plain and U-shape architecture, showing superior efficiency and competitive effectiveness. In addition to superior performance to DiT and other sub-quadratic-time diffusion models at 256 × 256 resolution, DiG demonstrates greater efficiency than these methods starting from a 512 resolution. Specifically, DiG-S/2 is 2.5× faster and saves 75.7% GPU memory compared to DiT-S/2 at a 1792 resolution. Additionally, DiG-XL/2 is 4.2× faster than the Mamba-based model at a 1024 resolution and 1.8× faster than DiT with FlashAttention-2 at a 2048 resolution.

Pixel grounding, encompassing tasks such as Referring Expression Segmentation (RES), has garnered considerable attention due to its immense potential for bridging the gap between vision and language modalities. However, advancements in this domain are currently constrained by limitations inherent in existing datasets, including limited object categories, insufficient textual diversity, and a scarcity of high-quality annotations. To mitigate these limitations, we introduce GroundingSuite, which comprises: (1) an automated data annotation framework leveraging multiple Vision-Language Model (VLM) agents; (2) a large-scale training dataset encompassing 9.56 million diverse referring expressions and their corresponding segmentations; and (3) a meticulously curated evaluation benchmark consisting of 3,800 images. The GroundingSuite training dataset facilitates substantial performance improvements, enabling models trained on it to achieve state-of-the-art results. Specifically, a cIoU of 68.9 on gRefCOCO and a gIoU of 55.3 on RefCOCOm. Moreover, the GroundingSuite annotation framework demonstrates superior efficiency compared to the current leading data annotation method, i.e., $4.5 \times$ faster than GLaMM.

Autoregressive (AR) models have reformulated image generation as next-token prediction, demonstrating remarkable potential and emerging as strong competitors to diffusion models. However, control-to-image generation, akin to ControlNet, remains largely unexplored within AR models. Although a natural approach, inspired by advancements in Large Language Models, is to tokenize control images into tokens and prefill them into the autoregressive model before decoding image tokens, it still falls short in generation quality compared to ControlNet and suffers from inefficiency. To this end, we introduce ControlAR, an efficient and effective framework for integrating spatial controls into autoregressive image generation models. Firstly, we explore control encoding for AR models and propose a lightweight control encoder to transform spatial inputs (e.g., canny edges or depth maps) into control tokens. Then ControlAR exploits the conditional decoding method to generate the next image token conditioned on the per-token fusion between control and image tokens, similar to positional encodings. Compared to prefilling tokens, using conditional decoding significantly strengthens the control capability of AR models but also maintains the model efficiency. Furthermore, the proposed ControlAR surprisingly empowers AR models with arbitrary-resolution image generation via conditional decoding and specific controls. Extensive experiments can demonstrate the controllability of the proposed ControlAR for the autoregressive control-to-image generation across diverse inputs, including edges, depths, and segmentation masks. Furthermore, both quantitative and qualitative results indicate that ControlAR surpasses previous state-of-the-art controllable diffusion models, e.g., ControlNet++.

Evaluating Large Language Models (LLMs) in open-ended scenarios is challenging because existing benchmarks and metrics can not measure them comprehensively. To address this problem, we propose to fine-tune LLMs as scalable judges (JudgeLM) to evaluate LLMs efficiently and effectively in open-ended benchmarks. We first propose a comprehensive, large-scale, high-quality dataset containing task seeds, LLMs-generated answers, and GPT-4-generated judgments for fine-tuning high-performance judges, as well as a new benchmark for evaluating the judges. We train JudgeLM at different scales from 7B, 13B, to 33B parameters, and conduct a systematic analysis of its capabilities and behaviors. We then analyze the key biases in fine-tuning LLM as a judge and consider them as position bias, knowledge bias, and format bias. To address these issues, JudgeLM introduces a bag of techniques including swap augmentation, reference support, and reference drop, which clearly enhance the judge's performance. JudgeLM obtains the state-of-the-art judge performance on both the existing PandaLM benchmark and our proposed new benchmark. Our JudgeLM is efficient and the JudgeLM-7B only needs 3 minutes to judge 5K samples with 8 A100 GPUs. JudgeLM obtains high agreement with the teacher judge, achieving an agreement exceeding 90% that even surpasses human-to-human agreement. JudgeLM also demonstrates extended capabilities in being judges of the single answer, multimodal models, multiple answers, multi-turn chat, etc.

Text-To-4D generation is rapidly developing and widely applied in various scenarios. However, existing methods often fail to incorporate adequate spatio-temporal modeling and prompt alignment within a unified framework, resulting in temporal inconsistencies, geometric distortions, or low-quality 4D content that deviates from the provided texts. Therefore, we propose STP4D, a novel approach that aims to integrate comprehensive spatio-temporal-prompt consistency modeling for high-quality text-to-4D generation. Specifically, STP4D employs three carefully designed modules: Time-Varying Prompt Embedding, Geometric Information Enhancement, and Temporal Extension Deformation, which collaborate to accomplish this goal. Furthermore, STP4D is among the first methods to exploit the Diffusion model to generate 4D Gaussians, combining the fine-grained modeling capabilities and the real-time rendering process of 4DGS with the rapid inference speed of the Diffusion model. Extensive experiments demonstrate that STP4D excels in generating high-fidelity 4D content with exceptional efficiency (approximately 4.6s per asset), surpassing existing methods in both quality and speed.

Reconstructing objects and extracting high-quality surfaces play a vital role in the real world. Current 4D representations show the ability to render high-quality novel views for dynamic objects, but cannot reconstruct high-quality meshes due to their implicit or geometrically inaccurate representations. In this paper, we propose a novel representation that can reconstruct accurate meshes from sparse image input, named Dynamic 2D Gaussians (D-2DGS). We adopt 2D Gaussians for basic geometry representation and use sparse-controlled points to capture the 2D Gaussian's deformation. By extracting the object mask from the rendered high-quality image and masking the rendered depth map, we remove floaters that are prone to occur during reconstruction and can extract high-quality dynamic mesh sequences of dynamic objects. Experiments demonstrate that our D-2DGS is outstanding in reconstructing detailed and smooth high-quality meshes from sparse inputs. The code is available at https://github.com/hustvl/Dynamic-2DGS.

Existing end-to-end autonomous driving (AD) algorithms typically follow the Imitation Learning (IL) paradigm, which faces challenges such as causal confusion and an open-loop gap. In this work, we propose RAD, a 3DGS-based closed-loop Reinforcement Learning (RL) framework for end-to-end Autonomous Driving. By leveraging 3DGS techniques, we construct a photorealistic digital replica of the real physical world, enabling the AD policy to extensively explore the state space and learn to handle out-of-distribution scenarios through large-scale trial and error. To enhance safety, we design specialized rewards to guide the policy in effectively responding to safety-critical events and understanding real-world causal relationships. To better align with human driving behavior, we incorporate IL into RL training as a regularization term. We introduce a closed-loop evaluation benchmark consisting of diverse, previously unseen 3DGS environments. Compared to IL-based methods, RAD achieves stronger performance in most closed-loop metrics, particularly exhibiting a 3× lower collision rate. Abundant closed-loop results are presented in the supplementary material. Code is available at https://github.com/hustvl/RAD for facilitating future research.

We present Genesis, a unified world model for joint generation of multi-view driving videos and LiDAR sequences with spatio-temporal and cross-modal consistency. Genesis employs a two-stage architecture that integrates a DiT-based video diffusion model with 3D-VAE encoding, and a BEV-represented LiDAR generator with NeRF-based rendering and adaptive sampling. Both modalities are directly coupled through a shared condition input, enabling coherent evolution across visual and geometric domains. To guide the generation with structured semantics, we introduce DataCrafter, a captioning module built on vision-language models that provides scene-level and instance-level captions. Extensive experiments on the nuScenes benchmark demonstrate that Genesis achieves state-of-the-art performance across video and LiDAR metrics (FVD 16.95, FID 4.24, Chamfer 0.611), and benefits downstream tasks including segmentation and 3D detection, validating the semantic fidelity and practical utility of the synthetic data.

Contrastive Language-Image Pretraining (CLIP) achieves strong generalization in vision-language tasks by aligning images and texts in a shared embedding space. However, recent findings show that CLIP-like models still underutilize fine-grained semantic signals in text, and this issue becomes even more pronounced when dealing with long and detailed captions. This stems from CLIP’s training objective, which optimizes only global image-text similarity and overlooks token-level supervision—limiting its ability to achieve fine-grained visual-text alignment. To address this, we propose SuperCLIP, a simple yet effective framework that augments contrastive learning with classification-based supervision. By adding only a lightweight linear layer to the vision encoder, SuperCLIP leverages token-level cues to enhance visual-textual alignment — with just a 0.077\% increase in total FLOPs, and no need for additional annotated data. Experiments show that SuperCLIP consistently improves zero-shot classification, image-text retrieval, and purely visual tasks. These gains hold regardless of whether the model is trained on original web data or rich re-captioned data, demonstrating SuperCLIP’s ability to recover textual supervision in both cases. Furthermore, SuperCLIP alleviates CLIP’s small-batch performance drop through classification-based supervision that avoids reliance on large batch sizes. Code and models will be made open source.

Exploring robust and efficient association methods has always been an important issue in multi-object tracking (MOT). Although existing tracking methods have achieved impressive performance, congestion and frequent occlusions still pose challenging problems in multi-object tracking. We reveal that performing sparse decomposition on dense scenes is a crucial step to enhance the performance of associating occluded targets. To this end, we propose a pseudo-depth estimation method for obtaining the relative depth of targets from 2D images. Secondly, we design a depth cascading matching (DCM) algorithm, which can use the obtained depth information to convert a dense target set into multiple sparse target subsets and perform data association on these sparse target subsets in order from near to far. By integrating the pseudo-depth method and the DCM strategy into the data association process, we propose a new tracker, called SparseTrack. SparseTrack provides a new perspective for solving the challenging crowded scene MOT problem. Only using IoU matching, SparseTrack achieves comparable performance with the state-of-the-art (SOTA) methods on the MOT17 and MOT20 benchmarks. Code and models are publicly available at https://github.com/hustvl/SparseTrack.

Four-dimensional Digital Subtraction Angiography (4D DSA) is a medical imaging technique that provides a series of 2D images captured at different stages and angles during the process of contrast agent filling blood vessels. It plays a significant role in the diagnosis of cerebrovascular diseases. Improving the rendering quality and speed under sparse sampling is important for observing the status and location of lesions. The current methods exhibit inadequate rendering quality in sparse views and suffer from slow rendering speed. To overcome these limitations, we propose TOGS, a Gaussian splatting method with opacity offset over time, which can effectively improve the rendering quality and speed of 4D DSA. We introduce an opacity offset table for each Gaussian to model the opacity offsets of the Gaussian, using these opacity-varying Gaussians to model the temporal variations in the radiance of the contrast agent. By interpolating the opacity offset table, the opacity variation of the Gaussian at different time points can be determined. This enables us to render the 2D DSA image at that specific moment. Additionally, we introduced a Smooth loss term in the loss function to mitigate overfitting issues that may arise in the model when dealing with sparse view scenarios. During the training phase, we randomly prune Gaussians, thereby reducing the storage overhead of the model. The experimental results demonstrate that compared to previous methods, this model achieves state-of-the-art render quality under the same number of training views. Additionally, it enables real-time rendering while maintaining low storage overhead.

Neural Radiances Fields (NeRF) and their extensions have shown great success in representing 3D scenes and synthesizing novel-view images. However, most NeRF methods take in low-dynamic-range (LDR) images, which may lose details, especially with nonuniform illumination. Some previous NeRF methods attempt to introduce high-dynamic-range (HDR) techniques but mainly target static scenes. To extend HDR NeRF methods to wider applications, we propose a dynamic HDR NeRF framework, named HDR-HexPlane, which can learn 3D scenes from dynamic 2D images captured with various exposures. A learnable exposure mapping function is constructed to obtain adaptive exposure values for each image. Based on the monotonically increasing prior, a camera response function is designed for stable learning. With the proposed model, high- quality novel-view images at any time point can be rendered with any desired exposure. We further construct a dataset containing multiple dynamic scenes captured with diverse exposures for evaluation. All the datasets and code are available at https://guanjunwu.github.io/HDR-HexPlane/.

Video instance segmentation on mobile devices is an important yet very challenging edge AI problem. It mainly suffers from (1) heavy computation and memory costs for frame-by-frame pixel-level instance perception and (2) complicated heuristics for tracking objects. To address these issues, we present MobileInst, a lightweight and mobile-friendly framework for video instance segmentation on mobile devices. Firstly, MobileInst adopts a mobile vision transformer to extract multi-level semantic features and presents an efficient query-based dual-transformer instance decoder for mask kernels and a semantic-enhanced mask decoder to generate instance segmentation per frame. Secondly, MobileInst exploits simple yet effective kernel reuse and kernel association to track objects for video instance segmentation. Further, we propose temporal query passing to enhance the tracking ability for kernels. We conduct experiments on COCO and YouTube-VIS datasets to demonstrate the superiority of MobileInst and evaluate the inference latency on one single CPU core of the Snapdragon 778G Mobile Platform, without other methods of acceleration. On the COCO dataset, MobileInst achieves 31.2 mask AP and 433 ms on the mobile CPU, which reduces the latency by 50% compared to the previous SOTA. For video instance segmentation, MobileInst achieves 35.0 AP and 30.1 AP on YouTube-VIS 2019 & 2021.

The You Only Look Once (YOLO) series of detectors have established themselves as efficient and practical tools. However, their reliance on predefined and trained object categories limits their applicability in open scenarios. Addressing this limitation, we introduce YOLO-World, an innovative approach that enhances YOLO with open-vocabulary detection capabilities through vision-language modeling and pre-training on large-scale datasets. Specifically, we propose a new Re-parameterizable Vision-Language Path Aggregation Network (RepVL-PAN) and region-text contrastive loss to facilitate the interaction between visual and linguistic information. Our method excels in detecting a wide range of objects in a zero-shot manner with high efficiency. On the challenging LVIS dataset, YOLO-World achieves 35.4 AP with 52.0 FPS on V100, which outperforms many state-of-the-art methods in terms of both accuracy and speed. Furthermore, the finetuned YOLO-World achieves remarkable performance on several downstream tasks, including object detection and open-vocabulary instance segmentation. Code and models are available at: https://github.com/AILab-eve/YOLO-World.

3D Semantic Scene Completion (SSC) has emerged as a nascent and pivotal undertaking in autonomous driving, aiming to predict the voxel occupancy within volumetric scenes. However, prevailing methodologies primarily focus on voxel-wise feature aggregation, while neglecting instance semantics and scene context. In this paper, we present a novel paradigm termed Symphonies (Scene-from-Insts), that delves into the integration of instance queries to orchestrate 2D-to-3D reconstruction and 3D scene modeling. Leveraging our proposed Serial Instance-Propagated Attentions, Symphonies dynamically encodes instance-centric semantics, facilitating intricate interactions between the image and volumetric domains. Simultaneously, Symphonies fosters holistic scene comprehension by capturing context through the efficient fusion of instance queries, alleviating geometric ambiguities such as occlusion and perspective errors through contextual scene reasoning. Experimental results demonstrate that Symphonies achieves state-of-the-art performance on the chal-lenging SemanticKITTI and SSCBench-KITTI-360 benchmarks, yielding remarkable mIoU scores of 15.04 and 18.58, respectively. These results showcase the promising advancements of our paradigm. The code for our method is available at https://github.com/hustvl/Symphonies.

Representing and rendering dynamic scenes has been an important but challenging task. Especially, to accurately model complex motions, high efficiency is usually hard to guarantee. To achieve real-time dynamic scene rendering while also enjoying high training and storage efficiency, we propose 4D Gaussian Splatting (4D-GS) as a holistic representation for dynamic scenes rather than applying 3D-GS for each individual frame. In 4D-GS, a novel explicit representation containing both 3D Gaussians and 4D neural voxels is proposed. A decomposed neural voxel encoding algorithm inspired by HexPlane is proposed to efficiently build Gaussian features from 4D neural voxels and then a lightweight MLP is applied to predict Gaussian deformations at novel timestamps. Our 4D-GS method achieves real-time rendering under high resolutions, 82 FPS at an 800x800 resolution on an RTX 3090 GPU while maintaining comparable or better quality than previous state- of-the-art methods. More demos and code are available at https://guanjunwu.github.io/4dgs/.

In recent times, the generation of 3D assets from text prompts has shown impressive results. Both 2D and 3D diffusion models can help generate decent 3D objects based on prompts. 3D diffusion models have good 3D consistency, but their quality and generalization are limited as trainable 3D data is expensive and hard to obtain. 2D diffusion models enjoy strong abilities of generalization and fine generation, but 3D consistency is hard to guarantee. This paper attempts to bridge the power from the two types of diffusion models via the recent explicit and efficient 3D Gaussian splatting representation. A fast 3D object gener-ation framework, named as GaussianDreamer, is proposed, where the 3D diffusion model provides priors for initial-ization and the 2D diffusion model enriches the geometry and appearance. Operations of noisy point growing and color perturbation are introduced to enhance the initialized Gaussians. Our GaussianDreamer can generate a high-quality 3D instance or 3D avatar within 15 minutes on one GPU, much faster than previous methods, while the generated instances can be directly rendered in real time. Demos and code are available at https://taoranyi.com/gaussiandreamer/.

Artificial intelligence has become a crucial tool for medical image analysis. As an advanced cerebral angiography technique, Digital Subtraction Angiography (DSA) poses a challenge where the radiation dose to humans is proportional to the image count. By reducing images and using AI interpolation instead, the radiation can be cut significantly. However, DSA images present more complex motion and structural features than natural scenes, making interpolation more challenging. We propose MoSt-DSA, the first work that uses deep learning for DSA frame interpolation. Unlike natural scene Video Frame Interpolation (VFI) methods that extract unclear or coarse-grained features, we devise a general module that models motion and structural context interactions between frames in an efficient full convolution manner by adjusting optimal context range and transforming contexts into linear functions. Benefiting from this, MoSt-DSA is also the first method that directly achieves any number of interpolations at any time steps with just one forward pass during both training and testing. We conduct extensive comparisons with 7 representative VFI models for interpolating 1 to 3 frames, MoSt-DSA demonstrates robust results across 470 DSA image sequences (each typically 152 images), with average SSIM over 0.93, average PSNR over 38 (standard deviations of less than 0.030 and 3.6, respectively), comprehensively achieving state-of-the-art performance in accuracy, speed, visual effect, and memory usage. Our code is available at https://github.com/ZyoungXu/MoSt-DSA.

Gait recognition, a growing field in biological recognition technology, utilizes distinct walking patterns for accurate individual identification. However, existing methods lack the incorporation of temporal information. To reach the full potential of gait recognition, we advocate for the consideration of temporal features at varying granularities and spans. This paper introduces a novel framework, GaitGS, which aggregates temporal features simultaneously in both granularity and span dimensions. Specifically, the Multi-Granularity Feature Extractor (MGFE) is designed to capture micro-motion and macro-motion information at fine and coarse levels respectively, while the Multi-Span Feature Extractor (MSFE) generates local and global temporal representations. Through extensive experiments on two datasets, our method demonstrates state-of-the-art performance, achieving Rank-1 accuracy of 98.2%, 96.5%, and 89.7% on CASIA-B under different conditions, and 97.6% on OU-MVLP. The source code will be available at https://github.com/Haijun-Xiong/GaitGS.

Recently the state space models (SSMs) with efficient hardware-aware designs, i.e., the Mamba deep learning model, have shown great potential for long sequence modeling. Meanwhile building efficient and generic vision backbones purely upon SSMs is an appealing direction. However, representing visual data is challenging for SSMs due to the position-sensitivity of visual data and the requirement of global context for visual understanding. In this paper, we show that the reliance on self-attention for visual representation learning is not necessary and propose a new generic vision backbone with bidirectional Mamba blocks (Vim), which marks the image sequences with position embeddings and compresses the visual representation with bidirectional state space models. On ImageNet classification, COCO object detection, and ADE20k semantic segmentation tasks, Vim achieves higher performance compared to well-established vision transformers like DeiT, while also demonstrating significantly improved computation & memory efficiency. For example, Vim is 2.8x faster than DeiT and saves 86.8% GPU memory when performing batch inference to extract features on images with a resolution of 1248x1248. The results demonstrate that Vim is capable of overcoming the computation & memory constraints on performing Transformer-style understanding for high-resolution images and it has great potential to be the next-generation backbone for vision foundation models.

Weakly-supervised visual recognition using inexact supervision is a critical yet challenging learning problem. It significantly reduces human labeling costs and traditionally relies on multi-instance learning and pseudo-labeling. This paper introduces WeakSAM and solves the weakly-supervised object detection (WSOD) and segmentation by utilizing the pre-learned world knowledge contained in a vision foundation model, i.e., the Segment Anything Model (SAM). WeakSAM addresses two critical limitations in traditional WSOD retraining, i.e., pseudo ground truth (PGT) incompleteness and noisy PGT instances, through adaptive PGT generation and Region of Interest (RoI) drop regularization. It also addresses the SAM's shortcomings of requiring human prompts and category unawareness in object detection and segmentation. Our results indicate that WeakSAM significantly surpasses previous state-of-the-art methods in WSOD and WSIS benchmarks with large margins, i.e. average improvements of 7.4% and 8.5%, respectively.

Diffusion Transformers (DiT) have attracted significant attention in research. However, they suffer from a slow convergence rate. In this paper, we aim to accelerate DiT training without any architectural modification. We identify the following issues in the training process: firstly, certain training strategies do not consistently perform well across different data. Secondly, the effectiveness of supervision at specific timesteps is limited. In response, we propose the following contributions: (1) We introduce a new perspective for interpreting the failure of the strategies. Specifically, we slightly extend the definition of Signal-to-Noise Ratio (SNR) and suggest observing the Probability Density Function (PDF) of SNR to understand the essence of the data robustness of the strategy. (2) We conduct numerous experiments and report over one hundred experimental results to empirically summarize a unified accelerating strategy from the perspective of PDF. (3) We develop a new supervision method that further accelerates the training process of DiT. Based on them, we propose FasterDiT, an exceedingly simple and practicable design strategy. With few lines of code modifications, it achieves 2.30 FID on ImageNet at 256x256 resolution with 1000 iterations, which is comparable to DiT (2.27 FID) but 7 times faster in training.

Feature pyramid representations have been widely adopted in the object detection literature for better handling of variations in scale, which provide abundant information from various spatial levels for classification and localization sub-tasks. We find that inter sub-task feature disentanglement and intra sub-task feature re-fusion are crucial for final prediction performance, but are hard to be achieved simultaneously considering the computational efficiency. We find this issue can be addressed by delicate module design. In this paper, we propose an Efficient Task-specific Feature Re-fusion (ETFR) module to mitigate the dilemma. ETFR disentangles inter sub-task features, reduces the output channels of multi-scale features based on their importance and re-fuses intra sub-task features via concatenation operation. As a plug-and-play module, ETFR can remarkably and consistently improve the well-established and highly-optimized object detection and instance segmentation methods, such as RetinaNet, FCOS, BlendMask and CondInst, with neglectable extra computation cost. Extensive experiments demonstrate that ETFR has good generalization ability on various changeling datasets, including COCO, LVIS and Cityscapes.

Labeling objects with pixel-wise segmentation requires a huge amount of human labor compared to bounding boxes. Most existing methods for weakly supervised instance segmentation focus on designing heuristic losses with priors from bounding boxes. While, we find that box-supervised methods can produce some fine segmentation masks and we wonder whether the detectors could learn from these fine masks while ignoring low-quality masks. To answer this question, we present BoxTeacher, an efficient and end-to-end training framework for high-performance weakly supervised instance segmentation, which leverages a sophisticated teacher to generate high-quality masks as pseudo labels. Considering the massive noisy masks hurt the training, we present a mask-aware confidence score to estimate the quality of pseudo masks, and propose the noiseaware pixel loss and noise-reduced affinity loss to adaptively optimize the student with pseudo masks. Extensive experiments can demonstrate effectiveness of the proposed BoxTeacher. Without bells and whistles, BoxTeacher remarkably achieves 35.0 mask AP and 36.5 mask AP with ResNet-50 and ResNet-101 respectively on the challenging COCO dataset, which outperforms the previous state-of-the-art methods by a significant margin and bridges the gap between box-supervised and mask-supervised methods.

We launch EVA, a vision-centric foundation model to Explore the limits of Visual representation at scAle using only publicly accessible data. EVA is a vanilla ViT pre-trained to reconstruct the masked out image-text aligned vision features conditioned on visible image patches. Via this pretext task, we can efficiently scale up EVA to one billion parameters, and sets new records on a broad range of representative vision downstream tasks, such as image recognition, video action recognition, object detection, instance segmentation and semantic segmentation without heavy supervised training. Moreover, we observe quantitative changes in scaling EVA result in qualitative changes in transfer learning performance that are not present in other models. For instance, EVA takes a great leap in the challenging large vocabulary instance segmentation task: our model achieves almost the same state-of-the-art performance on LVIS dataset with over a thousand categories and COCO dataset with only eighty categories. Beyond a pure vision encoder, EVA can also serve as a vision-centric, multi-modal pivot to connect images and text. We find initializing the vision tower of a giant CLIP from EVA can greatly stabilize the training and outperform the training from scratch counterpart with much fewer samples and less compute, providing a new direction for scaling up and accelerating the costly training of multi-modal foundation models.

Inspired by DETR variants, query-based end-to-end instance segmentation (QEIS) methods have recently outperformed CNN-based models on large-scale datasets. Yet they would lose efficacy when only a small amount of training data is available since it's hard for the crucial queries/kernels to learn localization and shape priors. To this end, this work offers a novel unsupervised pre-training solution for low-data regimes. Inspired by the recent success of the Prompting technique, we introduce a new pre-training method that boosts QEIS models by giving Saliency Prompt for queries/kernels. Our method contains three parts: 1) Saliency Masks Proposal is responsible for generating pseudo masks from unlabeled images based on the saliency mechanism. 2) Prompt-Kernel Matching transfers pseudo masks into prompts and injects the corresponding localization and shape priors to the best-matched kernels. 3) Kernel Supervision is applied to supply supervision at the kernel level for robust learning. From a practical perspective, our pre-training method helps QEIS models achieve a similar convergence speed and comparable performance with CNN-based models in low-data regimes. Experimental results show that our method significantly boosts several QEIS models on three datasets.11Code: https://github.com/lifuguan/saliency.prompt

Post-training quantization (PTQ) is a neural network compression technique that converts a full-precision model into a quantized model using lower-precision data types. Although it can help reduce the size and computational cost of deep neural networks, it can also introduce quantization noise and reduce prediction accuracy, especially in extremely low-bit settings. How to determine the appropriate quantization parameters (e.g., scaling factors and rounding of weights) is the main problem facing now. Existing methods attempt to determine these parameters by minimize the distance between features before and after quantization, but such an approach only considers local information and may not result in the most optimal quantization parameters. We analyze this issue and propose PD-Quant, a method that addresses this limitation by considering global information. It determines the quantization parameters by using the information of differences between network prediction before and after quantization. In addition, PD-Quant can alleviate the overfitting problem in PTQ caused by the small number of calibration sets by adjusting the distribution of activations. Experiments show that PD-Quant leads to better quantization parameters and improves the prediction accuracy of quantized models, especially in low-bit settings. For example, PD-Quant pushes the accuracy of ResNet-18 up to 53.14% and RegNetX-600MF up to 40.67% in weight 2-bit activation 2-bit. The code is released at https://github.com/hustv1/PD-Quant.

Both masked image modeling (MIM) and natural language supervision have facilitated the progress of transferable visual pre-training. In this work, we seek the synergy between two paradigms and study the emerging properties when MIM meets natural language supervision. To this end, we present a novel masked visual Reconstruction In Language semantic Space (RILS) pre-training framework, in which sentence representations, encoded by the text encoder, serve as prototypes to transform the vision-only signals into patch-sentence probabilities as semantically meaningful MIM reconstruction targets. The vision models can therefore capture useful components with structured information by predicting proper semantic of masked tokens. Better visual representations could, in turn, improve the text encoder via the image-text alignment objective, which is essential for the effective MIM target transformation. Extensive experimental results demonstrate that our method not only enjoys the best of previous MIM and CLIP but also achieves further improvements on various tasks due to their mutual benefits. RILS exhibits advanced transferability on downstream classification, detection, and segmentation, especially for low-shot regimes. Code is available at https://github.com/hustvl/RILS.

We present an approach to efficiently and effectively adapt a masked image modeling (MIM) pre-trained vanilla Vision Transformer (ViT) for object detection, which is based on our two novel observations: (i) A MIM pre-trained vanilla ViT encoder can work surprisingly well in the challenging object-level recognition scenario even with randomly sampled partial observations, e.g., only 25% ~ 50% of the input embeddings. (ii) In order to construct multi-scale representations for object detection from single-scale ViT, a randomly initialized compact convolutional stem supplants the pre-trained patchify stem, and its intermediate features can naturally serve as the higher resolution inputs of a feature pyramid network without further upsampling or other manipulations. While the pre-trained ViT is only regarded as the 3rd-stage of our detector’s backbone instead of the whole feature extractor. This naturally results in a ConvNet-ViT hybrid architecture. The proposed detector, named MimDet, enables a MIM pre-trained vanilla ViT to outperform leading hierarchical architectures such as Swin Transformer, MViTv2 and ConvNeXt on COCO object detection & instance segmentation, and achieves better results compared with the previous best adapted vanilla ViT detector using a more modest fine-tuning recipe while converging 2.8× faster. Code and pre-trained models are available at https://github.com/hustvl/MIMDet.

Autonomous driving requires a comprehensive understanding of the surrounding environment for reliable trajectory planning. Previous works rely on dense rasterized scene representation (e.g., agent occupancy and semantic map) to perform planning, which is computationally intensive and misses the instance-level structure information. In this paper, we propose VAD, an end-to-end vectorized paradigm for autonomous driving, which models the driving scene as a fully vectorized representation. The proposed vectorized paradigm has two significant advantages. On one hand, VAD exploits the vectorized agent motion and map elements as explicit instance-level planning constraints which effectively improves planning safety. On the other hand, VAD runs much faster than previous end-to-end planning methods by getting rid of computation-intensive rasterized representation and hand-designed post-processing steps. VAD achieves state-of-the-art end-to-end planning performance on the nuScenes dataset, outperforming the previous best method by a large margin. Our base model, VAD-Base, greatly reduces the average collision rate by 29.0% and runs 2.5× faster. Besides, a lightweight variant, VAD-Tiny, greatly improves the inference speed (up to 9.3×) while achieving comparable planning performance. We believe the excellent performance and the high efficiency of VAD are critical for the real-world deployment of an autonomous driving system. Code and models are available at https://github.com/hustvl/VAD for facilitating future research.

Category-level 6DoF object pose estimation intends to estimate the rotation, translation, and size of unseen objects. Many previous works use point clouds as a pre-learned shape prior to overcome intra-category variability. The shape prior is deformed to reconstruct instances’ point clouds in canonical space and to build dense 3D-3D correspondences between the observed and reconstructed point clouds. However, the pre-learned shape prior is not jointly optimized with estimation networks, and they are trained with a surrogate objective. We propose a novel 6D pose estimation network, named Query6DoF, based on a series of category-specific sparse queries that represent the prior shape. Each query represents a shape component, and these queries are learnable embeddings that can be optimized together with the estimation network according to the point cloud reconstruction loss, the normalized object coordinate loss, and the 6d pose estimation loss. Query6DoF adopts a deformation-and-matching paradigm with attention, where the queries dynamically extract features from regions of interest using the attention mechanism and then directly regress results. Furthermore, Query6DoF reduces computation overhead through the sparseness of the queries and the incorporation of a lightweight global information injection block. With the aforementioned design, Query6DoF achieves state-of-the-art (SOTA) pose estimation performance on the NOCS datasets. The source code and models are available at https://github.com/hustvl/Query6DoF.

In this paper, we propose a novel cross-modal distillation method, called TinyCLIP, for large-scale language-image pre-trained models. The method introduces two core techniques: affinity mimicking and weight inheritance. Affinity mimicking explores the interaction between modalities during distillation, enabling student models to mimic teachers’ behavior of learning cross-modal feature alignment in a visual-linguistic affinity space. Weight inheritance transmits the pre-trained weights from the teacher models to their student counterparts to improve distillation efficiency. Moreover, we extend the method into a multi-stage progressive distillation to mitigate the loss of informative weights during extreme compression. Comprehensive experiments demonstrate the efficacy of TinyCLIP, showing that it can reduce the size of the pre-trained CLIP ViT-B/32 by 50%, while maintaining comparable zero-shot performance. While aiming for comparable performance, distillation with weight inheritance can speed up the training by 1.4 - 7.8× compared to training from scratch. Moreover, our TinyCLIP ViT-8M/16, trained on YFCC-15M, achieves an impressive zero-shot top-1 accuracy of 41.1% on ImageNet, surpassing the original CLIP ViT-B/16 by 3.5% while utilizing only 8.9% parameters. Finally, we demonstrate the good transferability of TinyCLIP in various downstream tasks. Code and models will be open-sourced at aka.ms/tinyclip.

As the size of circuit designs continues to grow rapidly, artificial intelligence technologies are being extensively used in Electronic Design Automation (EDA) to assist with circuit design. Placement and routing are the most time-consuming parts of the physical design process, and how to quickly evaluate the placement has become a hot research topic. Prior works either transformed circuit designs into images using hand-crafted methods and then used Convolutional Neural Networks (CNN) to extract features, which are limited by the quality of the hand-crafted methods and could not achieve end-to-end training, or treated the circuit design as a graph structure and used Graph Neural Networks (GNN) to extract features, which require time-consuming preprocessing. In our work, we propose a novel perspective for circuit design by treating circuit components as point clouds and using Transformer-based point cloud perception methods to extract features from the circuit. This approach enables direct feature extraction from raw data without any preprocessing, allows for end-to-end training, and results in high performance. Experimental results show that our method achieves state-of-the-art performance in congestion prediction tasks on both the CircuitNet and ISPD2015 datasets, as well as in design rule check (DRC) violation prediction tasks on the CircuitNet dataset. Our method establishes a bridge between the relatively mature point cloud perception methods and the fast-developing EDA algorithms, enabling us to leverage more collective intelligence to solve this task. To facilitate the research of open EDA design, source codes and pre-trained models are released at https://github.com/hustvl/circuitformer.

Recent studies indicate that hierarchical Vision Transformer (ViT) with a macro architecture of interleaved non-overlapped window-based self-attention & shifted-window operation can achieve state-of-the-art performance in various visual recognition tasks, and challenges the ubiquitous convolutional neural networks (CNNs) using densely slid kernels. In most recently proposed hierarchical ViTs, self-attention is the de-facto standard for spatial information aggregation. In this paper, we question whether self-attention is the only choice for hierarchical ViT to attain strong performance, and study the effects of different kinds of cross-window communication methods. To this end, we replace self-attention layers with embarrassingly simple linear mapping layers, and the resulting proof-of-concept architecture termed TransLinear can achieve very strong performance in ImageNet-[Formula: see text] image recognition. Moreover, we find that TransLinear is able to leverage the ImageNet pre-trained weights and demonstrates competitive transfer learning properties on downstream dense prediction tasks such as object detection and instance segmentation. We also experiment with other alternatives to self-attention for content aggregation inside each non-overlapped window under different cross-window communication approaches. Our results reveal that the macro architecture, other than specific aggregation layers or cross-window communication mechanisms, is more responsible for hierarchical ViT's strong performance and is the real challenger to the ubiquitous CNN's dense sliding window paradigm.

Contextual information is vital in visual understanding problems, such as semantic segmentation and object detection. We propose a criss-cross network (CCNet) for obtaining full-image contextual information in a very effective and efficient way. Concretely, for each pixel, a novel criss-cross attention module harvests the contextual information of all the pixels on its criss-cross path. By taking a further recurrent operation, each pixel can finally capture the full-image dependencies. Besides, a category consistent loss is proposed to enforce the criss-cross attention module to produce more discriminative features. Overall, CCNet is with the following merits: 1) GPU memory friendly. Compared with the non-local block, the proposed recurrent criss-cross attention module requires 11× less GPU memory usage. 2) High computational efficiency. The recurrent criss-cross attention significantly reduces FLOPs by about 85 percent of the non-local block. 3) The state-of-the-art performance. We conduct extensive experiments on semantic segmentation benchmarks including Cityscapes, ADE20K, human parsing benchmark LIP, instance segmentation benchmark COCO, video segmentation benchmark CamVid. In particular, our CCNet achieves the mIoU scores of 81.9, 45.76 and 55.47 percent on the Cityscapes test set, the ADE20K validation set and the LIP validation set respectively, which are the new state-of-the-art results. The source codes are available at https://github.com/speedinghzl/CCNethttps://github.com/speedinghzl/CCNet.

We present VoxelTrack for multi-person 3D pose estimation and tracking from a few cameras which are separated by wide baselines. It employs a multi-branch network to jointly estimate 3D poses and re-identification (Re-ID) features for all people in the environment. In contrast to previous efforts which require to establish cross-view correspondence based on noisy 2D pose estimates, it directly estimates and tracks 3D poses from a 3D voxel-based representation constructed from multi-view images. We first discretize the 3D space by regular voxels and compute a feature vector for each voxel by averaging the body joint heatmaps that are inversely projected from all views. We estimate 3D poses from the voxel representation by predicting whether each voxel contains a particular body joint. Similarly, a Re-ID feature is computed for each voxel which is used to track the estimated 3D poses over time. The main advantage of the approach is that it avoids making any hard decisions based on individual images. The approach can robustly estimate and track 3D poses even when people are severely occluded in some cameras. It outperforms the state-of-the-art methods by a large margin on four public datasets including Shelf, Campus, Human3.6 M and CMU Panoptic.

Graph convolutional networks have been widely applied in skeleton-based gait recognition. A key challenge in this task is to distinguish the individual walking styles of different subjects across various views. Existing state-of-the-art methods employ uniform convolutions to extract features from diverse sequences and ignore the effects of viewpoint changes. To overcome these limitations, we propose a condition-adaptive graph (CAG) convolution network that can dynamically adapt to the specific attributes of each skeleton sequence and the corresponding view angle. In contrast to using fixed weights for all joints and sequences, we introduce a joint-specific filter learning (JSFL) module in the CAG method, which produces sequence-adaptive filters at the joint level. The adaptive filters capture fine-grained patterns that are unique to each joint, enabling the extraction of diverse spatial-temporal information about body parts. Additionally, we design a view-adaptive topology learning (VATL) module that generates adaptive graph topologies. These graph topologies are used to correlate the joints adaptively according to the specific view conditions. Thus, CAG can simultaneously adjust to various walking styles and viewpoints. Experiments on the two most widely used datasets (i.e., CASIA-B and OU-MVLP) show that CAG surpasses all previous skeleton-based methods. Moreover, the recognition performance can be enhanced by simply combining CAG with appearance-based methods, demonstrating the ability of CAG to provide useful complementary information.

Studying the inherent symmetry of data is of great importance in machine learning. Point cloud, the most important data format for 3D environmental perception, is naturally endowed with strong radial symmetry. In this work, we exploit this radial symmetry via a divide-and-conquer strategy to boost 3D perception performance and ease optimization. We propose Azimuth Normalization (AziNorm), which normalizes the point clouds along the radial direction and eliminates the variability brought by the difference of azimuth. AziNorm can be flexibly incorporated into most LiDAR-based perception methods. To validate its effectiveness and generalization ability, we apply AziNorm in both object detection and semantic segmentation. For detection, we integrate AziNorm into two representative detection methods, the one-stage SECOND detector and the state-of-the-art two-stage PV-RCNN detector. Experiments on Waymo Open Dataset demonstrate that AziNorm improves SECOND and PV-RCNN by 7.03 mAPH and 3.01 mAPH respectively. For segmentation, we integrate AziNorm into KPConv. On SemanticKitti dataset, AziNorm improves KPConv by 1.6/1.1 mIoU on val/test set. Besides, AziNorm remarkably improves data efficiency and accelerates convergence, reducing the requirement of data amounts or training epochs by an order of magnitude. SECOND w/ AziNorm can significantly outperform fully trained vanilla SECOND, even trained with only 10% data or 10% epochs. Code and models are available at https://github.com/hustvl/AziNorm.

In this paper, we propose a conceptually novel, efficient, and fully convolutional framework for real-time instance segmentation. Previously, most instance segmentation methods heavily rely on object detection and perform mask prediction based on bounding boxes or dense centers. In contrast, we propose a sparse set of instance activation maps, as a new object representation, to high-light informative regions for each foreground object. Then instance-level features are obtained by aggregating features according to the highlighted regions for recognition and segmentation. Moreover, based on bipartite matching, the instance activation maps can predict objects in a one-to-one style, thus avoiding non-maximum suppression (NMS) in post-processing. Owing to the simple yet effective designs with instance activation maps, SparseInst has extremely fast inference speed and achieves 40 FPS and 37.9 AP on the COCO benchmark, which significantly out-performs the counterparts in terms of speed and accuracy. Code and models are available at https://github.com/hustvl/SparseInst.

Transformers have offered a new methodology of designing neural networks for visual recognition. Compared to convolutional networks, Transformers enjoy the ability of referring to global features at each stage, yet the attention module brings higher computational overhead that obstructs the application of Transformers to process highresolution visual data. This paper aims to alleviate the conflict between efficiency and flexibility, for which we propose a specialized token for each region that serves as a messenger (MSG). Hence, by manipulating these MSG tokens, one can flexibly exchange visual information across regions and the computational complexity is reduced. We then integrate the MSG token into a multi-scale architecture named MSG-Transformer. In standard image classification and object detection, MSG-Transformer achieves competitive performance and the inference on both GPU and CPU is accelerated. Code is available at https://github.com/hustvl/MSG-Transformer.

Recently vision transformer has achieved tremendous success on image-level visual recognition tasks. To effectively and efficiently model the crucial temporal information within a video clip, we propose a Temporally Efficient Vision Transformer (TeViT) for video instance segmentation (VIS). Different from previous transformer-based VIS methods, TeViT is nearly convolution-free, which contains a transformer backbone and a query-based video instance segmentation head. In the backbone stage, we propose a nearly parameter-free messenger shift mechanism for early temporal context fusion. In the head stages, we propose a parameter-shared spatiotemporal query interaction mechanism to build the one-to-one correspondence between video instances and queries. Thus, TeViT fully utilizes both frame-level and instance-level temporal context information and obtains strong temporal modeling capacity with negligible extra computational cost. On three widely adopted VIS benchmarks, i.e., YouTube-VIS-2019, YouTube-VIS-2021, and OVIS, TeViT obtains state-of-the-art results and maintains high inference speed, e.g., 46.6 AP with 68.9 FPS on YouTube-VIS-2019. Code is available at https://github.com/hustvl/TeViT.

Although vision transformers (ViTs) have achieved great success in computer vision, the heavy computational cost hampers their applications to dense prediction tasks such as semantic segmentation on mobile devices. In this paper, we present a mobile-friendly architecture named Token Pyramid Vision Transformer (TopFormer). The proposed TopFormer takes Tokens from various scales as input to produce scale-aware semantic features, which are then in-Jected into the corresponding tokens to augment the representation. Experimental results demonstrate that our method significantly outperforms CNN- and ViT-based networks across several semantic segmentation datasets and achieves a good trade-off between accuracy and latency. On the ADE20K dataset, TopFormer achieves 5% higher accuracy in mIoU than MobileNetV3 with lower latency on an ARM-based mobile device. Furthermore, the tiny version of TopFormer achieves real-time inference on an ARM-based mobile device with competitive results. The code and models are available at: https://github.com/hustvl/TopFormer.

Neural radiance fields (NeRF) have shown great success in modeling 3D scenes and synthesizing novel-view images. However, most previous NeRF methods take much time to optimize one single scene. Explicit data structures, e.g. voxel features, show great potential to accelerate the training process. However, voxel features face two big challenges to be applied to dynamic scenes, i.e. modeling temporal information and capturing different scales of point motions. We propose a radiance field framework by representing scenes with time-aware voxel features, named as TiNeuVox. A tiny coordinate deformation network is introduced to model coarse motion trajectories and temporal information is further enhanced in the radiance network. A multi-distance interpolation method is proposed and applied on voxel features to model both small and large motions. Our framework significantly accelerates the optimization of dynamic radiance fields while maintaining high rendering quality. Empirical evaluation is performed on both synthetic and real scenes. Our TiNeuVox completes training with only 8 minutes and 8-MB storage cost while showing similar or even better rendering performance than previous dynamic NeRF methods. Code is available at https://github.com/hustvl/TiNeuVox.

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 .

Aggregating features in terms of different convolutional blocks or contextual embeddings has been proven to be an effective way to strengthen feature representations for semantic segmentation. However, most of the current popular network architectures tend to ignore the misalignment issues during the feature aggregation process caused by step-by-step downsampling operations and indiscriminate contextual information fusion. In this paper, we explore the principles in addressing such feature misalignment issues and inventively propose Feature-Aligned Segmentation Networks (AlignSeg). AlignSeg consists of two primary modules, i.e., the Aligned Feature Aggregation (AlignFA) module and the Aligned Context Modeling (AlignCM) module. First, AlignFA adopts a simple learnable interpolation strategy to learn transformation offsets of pixels, which can effectively relieve the feature misalignment issue caused by multi-resolution feature aggregation. Second, with the contextual embeddings in hand, AlignCM enables each pixel to choose private custom contextual information adaptively, making the contextual embeddings be better aligned. We validate the effectiveness of our AlignSeg network with extensive experiments on Cityscapes and ADE20K, achieving new state-of-the-art mIoU scores of 82.6 and 45.95 percent, respectively. Our source code is available at https://github.com/speedinghzl/AlignSeg.

Deep learning for computer vision has achieved remarkable results based on massive, diverse, and well-annotated training sets. However, it is difficult to collect defect datasets that cover all possible features, especially for small, weak defects. Therefore, in this article, a defect image generation method with controllable defect regions and strength is proposed. Regarded as image inpainting that uses a generative adversarial network, generated defect regions are controlled by using defect masks. Moreover, the defect direction vector is constructed in the latent variable space based on the feature continuity between defects and nondefects to control the defect strength, which enables a one-to-many correspondence between defect masks and images. Moreover, a defect attention loss is also designed to force the generation model to focus on the defect regions. Experimentally, our method yields generated images of better quality and diversity and thus significantly improves defect segmentation performance (an intersection over union of 63.20% and 61.86% on the Kolektor surface-defect and the metal hook defect datasets, respectively), especially for small, weak defects.

Humans have a strong class-agnostic object segmentation ability and can outline boundaries of unknown objects precisely, which motivates us to propose a box-supervised class-agnostic object segmentation (BoxCaseg) based solution for weakly-supervised instance segmentation. The BoxCaseg model is jointly trained using box-supervised images and salient images in a multi-task learning manner. The fine-annotated salient images provide class-agnostic and precise object localization guidance for box-supervised images. The object masks predicted by a pretrained BoxCaseg model are refined via a novel merged and dropped strategy as proxy ground truth to train a Mask R-CNN for weakly-supervised instance segmentation. Only using 7991 salient images, the weakly-supervised Mask R-CNN is on par with fully-supervised Mask R-CNN on PASCAL VOC and significantly outperforms previous state-of-the-art box-supervised instance segmentation methods on COCO. The source code, pretrained models and datasets are available at https://github.com/hustvl/BoxCaseg.

In this paper, we tackle the problem of human deocclusion which reasons about occluded segmentation masks and invisible appearance content of humans. In particular, a two-stage framework is proposed to estimate the invisible portions and recover the content inside. For the stage of mask completion, a stacked network structure is devised to refine inaccurate masks from a general instance segmentation model and predict integrated masks simultaneously. Additionally, the guidance from human parsing and typical pose masks are leveraged to bring prior information. For the stage of content recovery, a novel parsing guided attention module is applied to isolate body parts and capture context information across multiple scales. Besides, an Amodal Human Perception dataset (AHP) is collected to settle the task of human de-occlusion. AHP has advantages of providing annotations from real-world scenes and the number of humans is comparatively larger than other amodal perception datasets. Based on this dataset, experiments demonstrate that our method performs over the state-of-the-art techniques in both tasks of mask completion and content recovery. Our AHP dataset is available at https://sydney0zq.github.io/ahp/.

Instance segmentation on point clouds is a fundamental task in 3D scene perception. In this work, we propose a concise clustering-based framework named HAIS, which makes full use of spatial relation of points and point sets. Considering clustering-based methods may result in over-segmentation or under-segmentation, we introduce the hierarchical aggregation to progressively generate instance proposals, i.e., point aggregation for preliminarily clustering points to sets and set aggregation for generating complete instances from sets. Once the complete 3D instances are obtained, a sub-network of intra-instance prediction is adopted for noisy points filtering and mask quality scoring. HAIS is fast (only 410ms per frame on Titan X)) and does not require non-maximum suppression. It ranks 1st on the ScanNet v2 benchmark1, achieving the highest 69.9% AP50and surpassing previous state-of-the-art (SOTA) methods by a large margin. Besides, the SOTA results on the S3DIS dataset validate the good generalization ability. Code is available at https://github.com/hustvl/HAIS.

We present QueryInst, a new perspective for instance segmentation. QueryInst is a multi-stage end-to-end system that treats instances of interest as learnable queries, enabling query based object detectors, e.g., Sparse RCNN, to have strong instance segmentation performance. The attributes of instances such as categories, bounding boxes, instance masks, and instance association embeddings are represented by queries in a unified manner. In QueryInst, a query is shared by both detection and segmentation via dynamic convolutions and driven by parallellysupervised multi-stage learning. We conduct extensive experiments on three challenging benchmarks, i.e., COCO, CityScapes, and YouTube-VIS to evaluate the effectiveness of QueryInst in object detection, instance segmentation, and video instance segmentation tasks. For the first time, we demonstrate that a simple end-to-end query based framework can achieve the state-of-the-art performance in various instance-level recognition tasks. Code is available at https://github.com/hustvl/QueryInst.

Although gait recognition has drawn increasing research attention recently, it remains challenging to learn discriminative temporal representation since the silhouette differences are quite subtle in spatial domain. Inspired by the observation that humans can distinguish gaits of different subjects by adaptively focusing on temporal sequences with different time scales, we propose a context-sensitive temporal feature learning (CSTL) network in this paper, which aggregates temporal features in three scales to obtain motion representation according to the temporal contextual information. Specifically, CSTL introduces relation modeling among multi-scale features to evaluate feature importances, based on which network adaptively enhances more important scale and suppresses less important scale. Besides that, we propose a salient spatial feature learning (SSFL) module to tackle the misalignment problem caused by temporal operation, e.g., temporal convolution. SSFL recombines a frame of salient spatial features by extracting the most discriminative parts across the whole sequence. In this way, we achieve adaptive temporal learning and salient spatial mining simultaneously. Extensive experiments conducted on two datasets demonstrate the state-of-the-art performance. On CASIA-B dataset, we achieve rank-1 accuracies of 98.0%, 95.4% and 87.0% under normal walking, bag-carrying and coat-wearing conditions. On OU-MVLP dataset, we achieve rank-1 accuracy of 90.2%. The source code will be published at https://github.com/OliverHxh/CSTL.

Modeling temporal visual context across frames is critical for video instance segmentation (VIS) and other video understanding tasks. In this paper, we propose a fast on-line VIS model termed CrossVIS. For temporal information modeling in VIS, we present a novel crossover learning scheme that uses the instance feature in the current frame to pixel-wisely localize the same instance in other frames. Different from previous schemes, crossover learning does not require any additional network parameters for feature enhancement. By integrating with the instance segmentation loss, crossover learning enables efficient cross-frame instance-to-pixel relation learning and brings cost-free improvement during inference. Besides, a global balanced instance embedding branch is proposed for better and more stable online instance association. We conduct extensive experiments on three challenging VIS benchmarks, i.e., YouTube-VIS-2019, OVIS, and YouTube-VIS-2021 to evaluate our methods. CrossVIS achieves state-of-the-art online VIS performance and shows a decent trade-off between latency and accuracy. Code is available at https://github.com/hustvl/CrossVIS.

In this work, we propose a novel deep online correction (DOC) framework for monocular visual odometry. The whole pipeline has two stages: First, depth maps and initial poses are obtained from convolutional neural networks (CNNs) trained in self-supervised manners. Second, the poses predicted by CNNs are further improved by minimizing photometric errors via gradient updates of poses during inference phases. The benefits of our proposed method are twofold: 1) Different from online-learning methods, DOC does not need to calculate gradient propagation for parameters of CNNs. Thus, it saves more computation resources during inference phases. 2) Unlike hybrid methods that combine CNNs with traditional methods, DOC fully relies on deep learning (DL) frameworks. Though without complex back-end optimization modules, our method achieves outstanding performance with relative transform error (RTE) = 2.0% on KITTI Odometry benchmark for Seq. 09, which outperforms traditional monocular VO frameworks and is comparable to hybrid methods.

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.

Can Transformer perform $2\mathrm{D}$ object- and region-level recognition from a pure sequence-to-sequence perspective with minimal knowledge about the $2\mathrm{D}$ spatial structure? To answer this question, we present You Only Look at One Sequence (YOLOS), a series of object detection models based on the vanilla Vision Transformer with the fewest possible modifications, region priors, as well as inductive biases of the target task. We find that YOLOS pre-trained on the mid-sized ImageNet-$1k$ dataset only can already achieve quite competitive performance on the challenging COCO object detection benchmark, e.g., YOLOS-Base directly adopted from BERT-Base architecture can obtain $42.0$ box AP on COCO val. We also discuss the impacts as well as limitations of current pre-train schemes and model scaling strategies for Transformer in vision through YOLOS. Code and pre-trained models are available at https://github.com/hustvl/YOLOS.

Human pose estimation from image and video is a key task in many multimedia applications. Previous methods achieve great performance but rarely take efficiency into consideration, which makes it difficult to implement the networks on lightweight devices. Nowadays, real-time multimedia applications call for more efficient models for better interaction. Moreover, most deep neural networks for pose estimation directly reuse networks designed for image classification as the backbone, which are not optimized for the pose estimation task. In this paper, we propose an efficient framework for human pose estimation with two parts, an efficient backbone and an efficient head. By implementing a differentiable neural architecture search method, we customize the backbone network design for pose estimation, and reduce computational cost with negligible accuracy degradation. For the efficient head, we slim the transposed convolutions and propose a spatial information correction module to promote the performance of the final prediction. In experiments, we evaluate our networks on the MPII and COCO datasets. Our smallest model requires only 0.65 GFLOPs with 88.1% [email protected] on MPII and our large model needs only 2 GFLOPs while its accuracy is competitive with the state-of-the-art large model, HRNet, which takes 9.5 GFLOPs.

High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation through a subnetwork that is formed by connecting high-to-low resolution convolutions in series (e.g., ResNet, VGGNet), and then recover the high-resolution representation from the encoded low-resolution representation. Instead, our proposed network, named as High-Resolution Network (HRNet), maintains high-resolution representations through the whole process. There are two key characteristics: (i) Connect the high-to-low resolution convolution streams in parallel and (ii) repeatedly exchange the information across resolutions. The benefit is that the resulting representation is semantically richer and spatially more precise. We show the superiority of the proposed HRNet in a wide range of applications, including human pose estimation, semantic segmentation, and object detection, suggesting that the HRNet is a stronger backbone for computer vision problems. All the codes are available at https://github.com/HRNet.

Deep neural networks achieve remarkable performance in many computer vision tasks. Most state-of-the-art (SOTA) semantic segmentation and object detection approaches reuse neural network architectures designed for image classification as the backbone, commonly pre-trained on ImageNet. However, performance gains can be achieved by designing network architectures specifically for detection and segmentation, as shown by recent neural architecture search (NAS) research for detection and segmentation. One major challenge though is that ImageNet pre-training of the search space representation (a.k.a. super network) or the searched networks incurs huge computational cost. In this paper, we propose a Fast Network Adaptation (FNA++) method, which can adapt both the architecture and parameters of a seed network (e.g., an ImageNet pre-trained network) to become a network with different depths, widths, or kernel sizes via a parameter remapping technique, making it possible to use NAS for segmentation and detection tasks a lot more efficiently. In our experiments, we apply FNA++ on MobileNetV2 to obtain new networks for semantic segmentation, object detection, and human pose estimation that clearly outperform existing networks designed both manually and by NAS. We also implement FNA++ on ResNets and NAS networks, which demonstrates a great generalization ability. The total computation cost of FNA++ is significantly less than SOTA segmentation and detection NAS approaches: 1737× less than DPC, 6.8× less than Auto-DeepLab, and 8.0× less than DetNAS. A series of ablation studies are performed to demonstrate the effectiveness, and detailed analysis is provided for more insights into the working mechanism. Codes are available at https://github.com/JaminFong/FNA.

Few-shot learning is a challenging task that aims at training a classifier for unseen classes with only a few training examples. The main difficulty of few-shot learning lies in the lack of intra-class diversity within insufficient training samples. To alleviate this problem, we propose a novel generative framework, Diversity Transfer Network (DTN), that learns to transfer latent diversities from known categories and composite them with support features to generate diverse samples for novel categories in feature space. The learning problem of the sample generation (i.e., diversity transfer) is solved via minimizing an effective meta-classification loss in a single-stage network, instead of the generative loss in previous works. Besides, an organized auxiliary task co-training over known categories is proposed to stabilize the meta-training process of DTN. We perform extensive experiments and ablation studies on three datasets, i.e., miniImageNet, CIFAR100 and CUB. The results show that DTN, with single-stage training and faster convergence speed, obtains the state-of-the-art results among the feature generation based few-shot learning methods. Code and supplementary material are available at: https://github.com/Yuxin-CV/DTN.

Neural architecture search (NAS) has dramatically advanced the development of neural network design. We revisit the search space design in most previous NAS methods and find the number and widths of blocks are set manually. However, block counts and block widths determine the network scale (depth and width) and make a great influence on both the accuracy and the model cost (FLOPs/latency). In this paper, we propose to search block counts and block widths by designing a densely connected search space, i.e., DenseNAS. The new search space is represented as a dense super network, which is built upon our designed routing blocks. In the super network, routing blocks are densely connected and we search for the best path between them to derive the final architecture. We further propose a chained cost estimation algorithm to approximate the model cost during the search. Both the accuracy and model cost are optimized in DenseNAS. For experiments on the MobileNetV2-based search space, DenseNAS achieves 75.3% top-1 accuracy on ImageNet with only 361MB FLOPs and 17.9ms latency on a single TITAN-XP. The larger model searched by DenseNAS achieves 76.1% accuracy with only 479M FLOPs. DenseNAS further promotes the ImageNet classification accuracies of ResNet-18, -34 and -50-B by 1.5%, 0.5% and 0.3% with 200M, 600M and 680M FLOPs reduction respectively. The related code is available at https://github.com/JaminFong/DenseNAS.

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.

Compared with object detection in static images, object detection in videos is more challenging due to degraded image qualities. An effective way to address this problem is to exploit temporal contexts by linking the same object across video to form tubelets and aggregating classification scores in the tubelets. In this paper, we focus on obtaining high quality object linking results for better classification. Unlike previous methods that link objects by checking boxes between neighboring frames, we propose to link in the same frame. To achieve this goal, we extend prior methods in following aspects: (1) a cuboid proposal network that extracts spatio-temporal candidate cuboids which bound the movement of objects; (2) a short tubelet detection network that detects short tubelets in short video segments; (3) a short tubelet linking algorithm that links temporally-overlapping short tubelets to form long tubelets. Experiments on the ImageNet VID dataset show that our method outperforms both the static image detector and the previous state of the art. In particular, our method improves results by 8.8 percent over the static image detector for fast moving objects.

This article aims to improve deep-learning-based surface defect recognition. Owing to the insufficiency of the defect images in practical production lines and the high cost of labeling, it is difficult to obtain a sufficient defect data set in terms of diversity and quantity. A new generation method called surface defect-generation adversarial network (SDGAN), which employs generative adversarial networks (GANs), is proposed to generate defect images using a large number of defect-free images from industrial sites. Experiments show that the defect images generated by the SDGAN have better image quality and diversity than those generated by the state-of-the-art methods. The SDGAN is applied to expand the commutator cylinder surface defect image data sets with and without labels (referred to as the CCSD-L and CCSD-NL data sets, respectively). Regarding anomaly recognition, a 1.77% error rate and a 49.43% relative improvement (IMP) for the CCSD-NL defect data set are obtained. Regarding defect classification, a 0.74% error rate and a 57.47% IMP for the CCSD-L defect data set are achieved. Moreover, defect classification trained on the images augmented by the SDGAN is robust to uneven and poor lighting conditions. Note to Practitioners-This article proposes a method of defect image generation to address the lack of industrial defect images. Traditional defect recognition methods have two disadvantages: different types of defects require different algorithms and handcrafted features are deficient. Defect recognition using deep learning can solve the above problems. However, deep learning requires a plethora of images, and the number of industrial defect images cannot meet this requirement. We propose a new defect image-generation method called SDGAN to generate a defect image data set that balances diversity and authenticity. In practice, we employ a large number of defect-free images to generate a large number of defect images using our method to expand the industry defect-free image data set. Then, the augmented defect data set is used to build a deep-learning defect recognition model. Experiments show that the accuracy of defect recognition can be significantly improved by building a deep-learning defect recognition model using the augmented data set. Therefore, deep learning can achieve excellent performance in defect recognition with a limited number of defect images.

This paper presents an effective single network for hand keypoint detection, instead of relying on the frequently-used two-stage pipeline consisting of localizing the hand and detecting the key points. Our method trains a fully convolutional neural network in an end-to-end manner, based on a novelly proposed pose anchor network, which can be deemed as an extension of the region proposal network (RPN) in Faster Region-based convolutional network. Moreover, we generate our pose anchor in a data-driven way, i.e., a K-means cluster algorithm based on object keypoint similarity (OKS), instead of manually design. In this way, we can obtain multiple representative pose anchors with various gestures, angles, and scales. By introducing the pose anchor, we are capable of utilizing the prior knowledge of the hand structure, mitigating the problem of occlusion to some extent. We demonstrate the feasibility and effectiveness of our method with extensive experiments on the challenging large-scale multiview 3D hand pose dataset (LSM-HPD) and New Zealand Sign Language Dataset (NZSL).

Creating high-quality temporal action proposals is fundamental yet challenging for accurate action detection in untrimmed videos due to the complexity of the background and variation in actions' durations and magnitudes. In this paper, we propose a cascaded boundary network (CBN) to predict the action boundaries by considering the importance of precise boundary information to develop accurate action proposals. Specifically, the first stage of CBN locates the temporal boundaries by predicting the probability that each frame corresponds to an action, the start position, and the end position. A temporal convolutional network is used in this stage to capture short-term context information. Next, the predicted probabilities are forwarded to the second stage, in which a long short-term memory (LSTM) network is utilized for further refinement by exploiting the correlation between the predicted probabilities to capture long-term context information. Finally, we combine the results from both stages to produce a long- and short-term information fusion. The experiments on THUMOS14 and ActivityNet-1.3 show that CBN achieves state-of-the-art recall performance. The performance improvement is especially remarkable for a small average number (AN) of retrieved proposals; e.g., the average recall at AN=50 on THUMOS14 is improved from 37.46% to 43.06%. Further experiments are performed by introducing proposals generated by CBN into an existing action detection framework. CBN also achieves state-of-the-art average mAP@tIoU on the THUMOS14 detection benchmark.

Semantic image segmentation is an important yet unsolved problem. One of the major challenges is the large variability of the object scales. To tackle this scale problem, we propose a Scale-Adaptive Network (SAN) which consists of multiple branches with each one taking charge of the segmentation of the objects of a certain range of scales. Given an image, SAN first computes a dense scale map indicating the scale of each pixel which is automatically determined by the size of the enclosing object. Then the features of different branches are fused according to the scale map to generate the final segmentation map. To ensure that each branch indeed learns the features for a certain scale, we propose a scale-induced ground-truth map and enforce a scale-aware segmentation loss for the corresponding branch in addition to the final loss. Extensive experiments over the PASCAL-Person-Part, the PASCAL VOC 2012, and the Look into Person datasets demonstrate that our SAN can handle the large variability of the object scales and outperforms the state-of-the-art semantic segmentation methods.

Synthetic visual data refers to the data automatically rendered by the mature computer graphic algorithms. With the rapid development of these techniques, we can now collect photo-realistic synthetic images with accurate pixel-level annotations without much effort. However, due to the domain gaps between synthetic data and real data, in terms of not only visual appearance but also label distribution, directly applying models trained on synthetic images to real ones can hardly yield satisfactory performance. Since the collection of accurate labels for real images is very laborious and time-consuming, developing algorithms which can learn from synthetic images is of great significance. In this paper, we propose a novel framework, namely Active Pseudo-Labeling (APL), to reduce the domain gaps between synthetic images and real images. In APL framework, we first predict pseudo-labels for the unlabeled real images in the target domain by actively adapting the style of the real images to source domain. Specifically, the style of real images is adjusted via a novel task guided generative model, and then pseudo-labels are predicted for these actively adapted images. Lastly, we fine-tune the source-trained model in the pseudo-labeled target domain, which helps to fit the distribution of the real data. Experiments on both semantic segmentation and object detection tasks with several challenging benchmark data sets demonstrate the priority of our proposed method compared to the existing state-of-the-art approaches.

Accurate and rapid diagnosis of COVID-19 suspected cases plays a crucial role in timely quarantine and medical treatment. Developing a deep learning-based model for automatic COVID-19 diagnosis on chest CT is helpful to counter the outbreak of SARS-CoV-2. A weakly-supervised deep learning framework was developed using 3D CT volumes for COVID-19 classification and lesion localization. For each patient, the lung region was segmented using a pre-trained UNet; then the segmented 3D lung region was fed into a 3D deep neural network to predict the probability of COVID-19 infectious; the COVID-19 lesions are localized by combining the activation regions in the classification network and the unsupervised connected components. 499 CT volumes were used for training and 131 CT volumes were used for testing. Our algorithm obtained 0.959 ROC AUC and 0.976 PR AUC. When using a probability threshold of 0.5 to classify COVID-positive and COVID-negative, the algorithm obtained an accuracy of 0.901, a positive predictive value of 0.840 and a very high negative predictive value of 0.982. The algorithm took only 1.93 seconds to process a single patient's CT volume using a dedicated GPU. Our weakly-supervised deep learning model can accurately predict the COVID-19 infectious probability and discover lesion regions in chest CT without the need for annotating the lesions for training. The easily-trained and high-performance deep learning algorithm provides a fast way to identify COVID-19 patients, which is beneficial to control the outbreak of SARS-CoV-2. The developed deep learning software is available at https://github.com/sydney0zq/covid-19-detection.

State-of-the-art object detectors and trackers are developing fast. Trackers are in general more efficient than detectors but bear the risk of drifting. A question is hence raised – how to improve the accuracy of video object detection/tracking by utilizing the existing detectors and trackers within a given time budget? A baseline is frame skipping – detecting every N-th frames and tracking for the frames in between. This baseline, however, is suboptimal since the detection frequency should depend on the tracking quality. To this end, we propose a scheduler network, which determines to detect or track at a certain frame, as a generalization of Siamese trackers. Although being light-weight and simple in structure, the scheduler network is more effective than the frame skipping baselines and flow-based approaches, as validated on ImageNet VID dataset in video object detection/tracking.

Neural Architecture Search (NAS) is an important yet challenging task in network design due to its high computational consumption. To address this issue, we propose the Reinforced Evolutionary Neural Architecture Search (RENAS), which is an evolutionary method with reinforced mutation for NAS. Our method integrates reinforced mutation into an evolution algorithm for neural architecture exploration, in which a mutation controller is introduced to learn the effects of slight modifications and make mutation actions. The reinforced mutation controller guides the model population to evolve efficiently. Furthermore, as child models can inherit parameters from their parents during evolution, our method requires very limited computational resources. In experiments, we conduct the proposed search method on CIFAR-10 and obtain a powerful network architecture, RENASNet. This architecture achieves a competitive result on CIFAR-10. The explored network architecture is transferable to ImageNet and achieves a new state-of-the-art accuracy, i.e., 75.7% top-1 accuracy with 5.36M parameters on mobile ImageNet. We further test its performance on semantic segmentation with DeepLabv3 on the PASCAL VOC. RENASNet outperforms MobileNet-v1, MobileNet-v2 and NASNet. It achieves 75.83% mIOU without being pretrained on COCO.

Letting a deep network be aware of the quality of its own predictions is an interesting yet important problem. In the task of instance segmentation, the confidence of instance classification is used as mask quality score in most instance segmentation frameworks. However, the mask quality, quantified as the IoU between the instance mask and its ground truth, is usually not well correlated with classification score. In this paper, we study this problem and propose Mask Scoring R-CNN which contains a network block to learn the quality of the predicted instance masks. The proposed network block takes the instance feature and the corresponding predicted mask together to regress the mask IoU. The mask scoring strategy calibrates the misalignment between mask quality and mask score, and improves instance segmentation performance by prioritizing more accurate mask predictions during COCO AP evaluation. By extensive evaluations on the COCO dataset, Mask Scoring R-CNN brings consistent and noticeable gain with different models and outperforms the state-of-the-art Mask R-CNN. We hope our simple and effective approach will provide a new direction for improving instance segmentation. The source code of our method is available at \url{https://github.com/zjhuang22/maskscoring_rcnn}.

Visual object recognition under situations in which the direct line-of-sight is blocked, such as when it is occluded around the corner, is of practical importance in a wide range of applications. With coherent illumination, the light scattered from diffusive walls forms speckle patterns that contain information of the hidden object. It is possible to realize non-line-of-sight (NLOS) recognition with these speckle patterns. We introduce a novel approach based on speckle pattern recognition with deep neural network, which is simpler and more robust than other NLOS recognition methods. Simulations and experiments are performed to verify the feasibility and performance of this approach.

Full-image dependencies provide useful contextual information to benefit visual understanding problems. In this work, we propose a Criss-Cross Network (CCNet) for obtaining such contextual information in a more effective and efficient way. Concretely, for each pixel, a novel criss-cross attention module in CCNet harvests the contextual information of all the pixels on its criss-cross path. By taking a further recurrent operation, each pixel can finally capture the full-image dependencies from all pixels. Overall, CCNet is with the following merits: 1) GPU memory friendly. Compared with the non-local block, the proposed recurrent criss-cross attention module requires 11x less GPU memory usage. 2) High computational efficiency. The recurrent criss-cross attention significantly reduces FLOPs by about 85% of the non-local block in computing full-image dependencies. 3) The state-of-the-art performance. We conduct extensive experiments on popular semantic segmentation benchmarks including Cityscapes, ADE20K, and instance segmentation benchmark COCO. In particular, our CCNet achieves the mIoU score of 81.4 and 45.22 on Cityscapes test set and ADE20K validation set, respectively, which are the new state-of-the-art results. The source code is available at https://github.com/speedinghzl/CCNet.

In this paper, we investigate the problem of resource-constrained object detection using deep learning, which is a challenging problem in real-world applications. To address this problem, we propose a single-shot detector with multiple inference paths based on a multi-scale DenseNet. Experiments are carried out on the PASCAL VOC and COCO datasets, and results show that, with significant computation reduction, our detection network obtains comparable performance corresponding to the single-shot detectors, such as YOLO and SSD.

A challenging aspect of scene text recognition is to handle text with distortions or irregular layout. In particular, perspective text and curved text are common in natural scenes and are difficult to recognize. In this work, we introduce ASTER, an end-to-end neural network model that comprises a rectification network and a recognition network. The rectification network adaptively transforms an input image into a new one, rectifying the text in it. It is powered by a flexible Thin-Plate Spline transformation which handles a variety of text irregularities and is trained without human annotations. The recognition network is an attentional sequence-to-sequence model that predicts a character sequence directly from the rectified image. The whole model is trained end to end, requiring only images and their groundtruth text. Through extensive experiments, we verify the effectiveness of the rectification and demonstrate the state-of-the-art recognition performance of ASTER. Furthermore, we demonstrate that ASTER is a powerful component in end-to-end recognition systems, for its ability to enhance the detector.

Despite the great success of convolutional neural networks (CNNs) for the image classification task on data sets such as Cifar and ImageNet, CNN's representation power is still somewhat limited in dealing with images that have a large variation in size and clutter, where Fisher vector (FV) has shown to be an effective encoding strategy. FV encodes an image by aggregating local descriptors with a universal generative Gaussian mixture model (GMM). FV, however, has limited learning capability and its parameters are mostly fixed after constructing the codebook. To combine together the best of the two worlds, we propose in this brief a neural network structure with FV layer being part of an end-to-end trainable system that is differentiable; we name our network FisherNet that is learnable using back propagation. Our proposed FisherNet combines CNN training and FV encoding in a single end-to-end structure. We observe a clear advantage of FisherNet over plain CNN and standard FV in terms of both classification accuracy and computational efficiency on the challenging PASCAL visual object classes object classification and emotion image classification tasks.

End-to-end optimization of multi-instance learning (MIL) using neural networks is an important problem with many applications, in which a core issue is how to design a permutation-invariant pooling function without losing much instance-level information. Inspired by the dynamic routing in recent capsule networks, we propose a novel dynamic pooling function for MIL. It is an adaptive scheme for both key instance selection and modeling the contextual information among instances in a bag. The dynamic pooling iteratively updates the instance contribution to its bag. It is permutation-invariant and can interpret instance-to-bag relationship. The proposed dynamic pooling based multi-instance neural network has been validated on many MIL tasks and outperforms other MIL methods.

This paper studies the problem of learning image semantic segmentation networks only using image-level labels as supervision, which is important since it can significantly reduce human annotation efforts. Recent state-of-the-art methods on this problem first infer the sparse and discriminative regions for each object class using a deep classification network, then train semantic a segmentation network using the discriminative regions as supervision. Inspired by the traditional image segmentation methods of seeded region growing, we propose to train a semantic segmentation network starting from the discriminative regions and progressively increase the pixel-level supervision using by seeded region growing. The seeded region growing module is integrated in a deep segmentation network and can benefit from deep features. Different from conventional deep networks which have fixed/static labels, the proposed weakly-supervised network generates new labels using the contextual information within an image. The proposed method significantly outperforms the weakly-supervised semantic segmentation methods using static labels, and obtains the state-of-the-art performance, which are 63.2% mIoU score on the PASCAL VOC 2012 test set and 26.0% mIoU score on the COCO dataset.

Object detection is a core problem in computer vision and pattern recognition. In this paper, we study the problem of learning an effective object detector using weakly-annotated images (i.e., only the image level annotation is given) and a small proportion of fully-annotated images (i.e., bounding box level annotation is given) with curriculum learning. Our method is built upon Faster R-CNN. Different from previous weakly-supervised object detectors which rely on hand-craft object proposals, the proposed method learns a region proposal network using weakly- and semi-supervised training data. And the weakly-labeled images are fed into the deep network in a meaningful order which illustrates from easy to gradually more complex examples with curriculum learning. We name the Faster R-CNN trained using Weakly- And Semi-Supervised data with Curriculum Learning as WASSCL R-CNN. The WASSCL R-CNN is validated on the PASCAL VOC 2007 benchmark, and obtains 90% of a fully-supervised Faster R-CNN's performance (measured using mAP) with only 15% of fully-supervised annotations together with weak supervision. The results show that the proposed learning framework can significantly reduce the labeling efforts for obtaining reliable object detectors.

The rapid popularity of mobile devices especially smart phones has changed human life style greatly. In this paper, we examine the consumer behaviors on several e-commerce platforms based on a large-scale dataset of mobile internet access records for about 3.5 months from a major telecom operator in China, which covers 126,388 users from Shanghai. We provide a preliminary study on users' daily and periodic online shopping behaviors, as well as the influence of special online shopping events and gender factors. These findings may be exploited by e-commerce providers e.g., for developing personalized recommendation systems to improve their service quality and profit.

The proliferation of mobile devices especially smart phones brings remarkable opportunities for both industry and academia. In particular, the massive data generated from users» usage logs provide the possibilities for stakeholders to know better about consumer behaviors with the aid of data mining. In this paper, we examine the consumer behaviors across multiple platforms based on a large-scale mobile Internet dataset from a major telecom operator, which covers 9.8 million users from two regions among which 1.4 million users have visited e-commerce platforms within one week of our study. We make several interesting observations and examine users» cultural differences from different regions. Our analysis shows among the multiple e-commerce platforms available, most mobile users are loyal to their favorable sites; people (60%) tend to make quick decisions to buy something online, which usually takes less than half an hour. Furthermore, we find that people in residential areas are much easier to perform purchases than in business districts and purchases take place during non-work time. Meanwhile, people with medium socioeconomic status like browsing and purchasing on e-commerce platforms, while people with high and low socioeconomic status are much easier to conduct purchases online. We also show the predictability of cross-platform shopping behaviors with extensive experiments on the basis of our observed data. Our discoveries could be a good guide for e-commerce future strategy making.

Summary Addressing service control factors, rapid manufacturing environment change, difficulty of resource allocation evaluation, resource optimization of 3D cloud printing service in a cloud manufacturing environment, and other characteristics, this paper proposes an evaluation indicator system of innovative new product development 3D printing order task execution. The evaluation indicator has eight dimensional components, including Time (T), Quality of Service (Q), Matching (Mat), Reliability (R), Flexibility (Flex), Cost (C), Fault tolerance (Ft), and Satisfaction (Sa). It constructs a type of optimal selection model based on a Multi‐Agent 3D Cloud Printing Service Quality Evaluation and a framework of cloud service evaluation of an AHP‐TOPSIS evaluation model based on Pareto optimization, and it designs an algorithm involving hybrid multi‐objective particle swarm optimization (PSO) based on the Baldwin Effect Model. In addition, this paper verifies the effectiveness of the algorithm through an example and offers a case study designed to test its feasibility and effectiveness.

This paper presents an end-to-end trainable fast scene text detector, named TextBoxes, which detects scene text with both high accuracy and efficiency in a single network forward pass, involving no post-process except for a standard non-maximum suppression. TextBoxes outperforms competing methods in terms of text localization accuracy and is much faster, taking only 0.09s per image in a fast implementation. Furthermore, combined with a text recognizer, TextBoxes significantly outperforms state-of-the-art approaches on word spotting and end-to-end text recognition tasks.

Of late, weakly supervised object detection is with great importance in object recognition. Based on deep learning, weakly supervised detectors have achieved many promising results. However, compared with fully supervised detection, it is more challenging to train deep network based detectors in a weakly supervised manner. Here we formulate weakly supervised detection as a Multiple Instance Learning (MIL) problem, where instance classifiers (object detectors) are put into the network as hidden nodes. We propose a novel online instance classifier refinement algorithm to integrate MIL and the instance classifier refinement procedure into a single deep network, and train the network end-to-end with only image-level supervision, i.e., without object location information. More precisely, instance labels inferred from weak supervision are propagated to their spatially overlapped instances to refine instance classifier online. The iterative instance classifier refinement procedure is implemented using multiple streams in deep network, where each stream supervises its latter stream. Weakly supervised object detection experiments are carried out on the challenging PASCAL VOC 2007 and 2012 benchmarks. We obtain 47% mAP on VOC 2007 that significantly outperforms the previous state-of-the-art.

Edge and surface are two fundamental visual elements of an object. The majority of existing object proposal approaches utilize edge or edge-like cues to rank candidates, while we consider that the surface cue containing the 3D characteristic of objects should be captured effectively for proposals, which has been rarely discussed before. In this paper, an object-level proposal model is presented, which constructs an occlusion-based objectness taking the surface cue into account. Specifically, the better detection of occlusion edges is focused on to enrich the surface cue into proposals, namely, the occlusion-dominated fusion and normalization criterion are designed to obtain the approximately overall contour information, to enhance the occlusion edge map at utmost and thus boost proposals. Experimental results on the PASCAL VOC 2007 and MS COCO 2014 dataset demonstrate the effectiveness of our approach, which achieves around 6% improvement on the average recall than Edge Boxes at 1000 proposals and also leads to a modest gain on the performance of object detection.

This paper presents a novel and robust descriptor, depth-projection-map-based bag of contour fragments, which is applied to extraction of hand shape and structure information from depth maps. Our method projects depth maps onto three orthogonal planes to generate the depth projection maps. Then, the bag of contour fragment descriptors are extracted from the three depth projection maps and concatenated as a final shape representation of the original depth data. A support vector machine with a linear kernel is used as a shape classifier. The proposed description method is evaluated on three public datasets, as well as a new and more challenging dataset for hand gesture recognition. Results demonstrate that the proposed method significantly outperforms the previous methods on all tested datasets for both static digit recognition and letter gesture recognition. For the challenging HUST-ASL dataset, in particular, the proposed method improves on the previous state-of-the-art methods from 40.1% to 64.6%.

Finding an effective and efficient representation is very important for image classification. The most common approach is to extract a set of local descriptors, and then aggregate them into a high-dimensional, more semantic feature vector, like unsupervised bag-of-features and weakly supervised part-based models. The latter one is usually more discriminative than the former due to the use of information from image labels. In this paper, we propose a weakly supervised strategy that using multi-instance learning (MIL) to learn discriminative patterns for image representation. Specially, we extend traditional multi-instance methods to explicitly learn more than one patterns in positive class, and find the "most positive" instance for each pattern. Furthermore, as the positiveness of instance is treated as a continuous variable, we can use stochastic gradient decent to maximize the margin between different patterns meanwhile considering MIL constraints. To make the learned patterns more discriminative, local descriptors extracted by deep convolutional neural networks are chosen instead of hand-crafted descriptors. Some experimental results are reported on several widely used benchmarks (Action 40, Caltech 101, Scene 15, MIT-indoor, SUN 397), showing that our method can achieve very remarkable performance.

Recognizing text in natural images is a challenging task with many unsolved problems. Different from those in documents, words in natural images often possess irregular shapes, which are caused by perspective distortion, curved character placement, etc. We propose RARE (Robust text recognizer with Automatic REctification), a recognition model that is robust to irregular text. RARE is a speciallydesigned deep neural network, which consists of a Spatial Transformer Network (STN) and a Sequence Recognition Network (SRN). In testing, an image is firstly rectified via a predicted Thin-Plate-Spline (TPS) transformation, into a more "readable" image for the following SRN, which recognizes text through a sequence recognition approach. We show that the model is able to recognize several types of irregular text, including perspective text and curved text. RARE is end-to-end trainable, requiring only images and associated text labels, making it convenient to train and deploy the model in practical systems. State-of-the-art or highly-competitive performance achieved on several benchmarks well demonstrates the effectiveness of the proposed model.

Contour detection serves as the basis of a variety of computer vision tasks such as image segmentation and object recognition. The mainstream works to address this problem focus on designing engineered gradient features. In this work, we show that contour detection accuracy can be improved by instead making the use of the deep features learned from convolutional neural networks (CNNs). While rather than using the networks as a blackbox feature extractor, we customize the training strategy by partitioning contour (positive) data into subclasses and fitting each subclass by different model parameters. A new loss function, named positive-sharing loss, in which each subclass shares the loss for the whole positive class, is proposed to learn the parameters. Compared to the sofmax loss function, the proposed one, introduces an extra regularizer to emphasizes the losses for the positive and negative classes, which facilitates to explore more discriminative features. Our experimental results demonstrate that learned deep features can achieve top performance on Berkeley Segmentation Dataset and Benchmark (BSDS500) and obtain competitive cross dataset generalization result on the NYUD dataset.

Multiple-instance learning (MIL) has served as an important tool for a wide range of vision applications, for instance, image classification, object detection, and visual tracking. In this paper, we propose a novel method to solve the classical MIL problem, named relaxed multiple-instance SVM (RMI-SVM). We treat the positiveness of instance as a continuous variable, use Noisy-OR model to enforce the MIL constraints, and optimize them jointly in a unified framework. The optimization problem can be efficiently solved using stochastic gradient decent. The extensive experiments demonstrate that RMI-SVM consistently achieves superior performance on various benchmarks for MIL. Moreover, we simply applied RMI-SVM to a challenging vision task, common object discovery. The state-of-the arts results of object discovery on PASCAL VOC datasets further confirm the advantages of the proposed method.

In this paper, we study the problem of recognizing car brands in surveillance videos, cast it as an image classification problem, and propose a novel multiple instance learning method, named Spatially Coherent Discriminative Pattern Learning, to discover the most discriminative patterns in car images. The learned discriminative patterns can effectively distinguish cars of different brands with high accuracy and efficiency. The experimental results demonstrate that our method is significantly superior to recent image classification methods on this problem. The proposed method is able to deliver an end-to-end real-time car recognition system for video surveillance. Moreover, we construct a large and challenging car image data set, consisting of 37 195 real-world car images from 30 brands, which could serve as a standard benchmark in this field and be used in various related research communities.

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.

The state-of-the-art SIFT flow has been widely adopted for the general image matching task, especially in dealing with image pairs from similar scenes but with different object configurations. However, the way in which the dense SIFT features are computed at a fixed scale in the SIFT flow method limits its capability of dealing with scenes of large scale changes. In this paper, we propose a simple, intuitive, and very effective approach, Scale-Space SIFT flow, to deal with the large scale differences in different image locations. We introduce a scale field to the SIFT flow function to automatically explore the scale deformations. Our approach achieves similar performance as the SIFT flow method on general natural scenes but obtains significant improvement on the images with large scale differences. Compared with a recent method that addresses the similar problem, our approach shows its clear advantage being more effective, and significantly less demanding in memory and time requirement.

This letter examines the problem of robust subspace discovery from input data samples (instances) in the presence of overwhelming outliers and corruptions. A typical example is the case where we are given a set of images; each image contains, for example, a face at an unknown location of an unknown size; our goal is to identify or detect the face in the image and simultaneously learn its model. We employ a simple generative subspace model and propose a new formulation to simultaneously infer the label information and learn the model using low-rank optimization. Solving this problem enables us to simultaneously identify the ownership of instances to the subspace and learn the corresponding subspace model. We give an efficient and effective algorithm based on the alternating direction method of multipliers and provide extensive simulations and experiments to verify the effectiveness of our method. The proposed scheme can also be used to tackle many related high-dimensional combinatorial selection problems.

In this paper, a learning-based shape descriptor for shape matching is demonstrated. Formulated in a bag-of-words like framework, the proposed method summarizes the local features extracted from certain shape to generate a integrated representation. It contributes to the speed-up of shape matching, since the distance metric in the vector space analysis can be directly applied to compare the constructed global descriptors, eliminating the time consuming stage of local feature matching. Similar to the philosophy in spatial pyramid matching, a strategy for feature division is applied in the phase of encoded feature pooling and vocabulary learning, which helps to construct a more discriminative descriptor incorporating both global and local information. Also, a local contour-based feature extraction method is designed for 2D shapes, while significant properties of the local contours are inspected for the design of feature division rules. The designed local feature extraction method and the feature division rules manage to reduce the variances of shape representation due to the changes in rotation. In addition to 2D shape, we also present a simple and natural method to extend the proposed method to the scenario of 3D shape representation. The proposed shape descriptor is validated on several benchmark data sets for evaluating 2D and 3D shape matching algorithms, and it is observed that the investigated shape descriptor maintains superior discriminative power as well as high time efficiency.

This paper makes use of locality-constrained linear coding (LLC) in a two-layer image representation framework for traffic sign recognition. As a multi-category classification problem with unbalanced frequencies and variations, many machine learning approaches have been adopted with some low level features for traffic sign recognition. To the best of our knowledge, this is the first method using coding features for traffic sign recognition. First, we extract features(dense SIFT features, HOG features and LBP features) and encode them with a k-means generated codebook and LLC. Second, each traffic sign image is represented by the features generated by spatial pyramid matching (SPM). Then, all the image representations from each kind of features are concatenated together as the final image representation. Finally, we show that a linear SVM classifier trained with this image representation can achieve the state-of-the-art recognition rate of 99.67% on the well-known German Traffic Sign Recognition Benchmark.

Dictionary learning has became an increasingly important task in machine learning, as it is fundamental to the representation problem. A number of emerging techniques specifically include a codebook learning step, in which a critical knowledge abstraction process is carried out. Existing approaches in dictionary (codebook) learning are either generative (unsupervised e.g. k-means) or discriminative (supervised e.g. extremely randomized forests). In this paper, we propose a multiple instance learning (MIL) strategy (along the line of weakly supervised learning) for dictionary learning. Each code is represented by a classifier, such as a linear SVM, which naturally performs metric fusion for multi-channel features. We design a formulation to simultaneously learn mixtures of codes by maximizing classification margins in MIL. State-of-the-art results are observed in image classification benchmarks based on the learned codebooks, which observe both compactness and effectiveness.

We propose a novel shape model for object detection called Fan Shape Model (FSM). We model contour sample points as rays of final length emanating for a reference point. As in folding fan, its slats, which we call rays, are very flexible. This flexibility allows FSM to tolerate large shape variance. However, the order and the adjacency relation of the slats stay invariant during fan deformation, since the slats are connected with a thin fabric. In analogy, we enforce the order and adjacency relation of the rays to stay invariant during the deformation. Therefore, FSM preserves discriminative power while allowing for a substantial shape deformation. FSM allows also for precise scale estimation during object detection. Thus, there is not need to scale the shape model or image in order to perform object detection. Another advantage of FSM is the fact that it can be applied directly to edge images, since it does not require any linking of edge pixels to edge fragments (contours).

In this paper, we presents a new method to encode the spatial information of local image features, which is a natural extension of Shape Context (SC), so we call it Feature Context (FC). Given a position in a image, SC computes histogram of other points belonging to the target binary shape based on their distances and angles to the position. The value of each histogram bin of SC is the number of the shape points in the region assigned to the bin. Thus, SC requires knowing the location of the points of the target shape. In other words, an image point can have only two labels, it belongs to the shape or not. In contrast, FC can be applied to the whole image without knowing the location of the target shape in the image. Each image point can have multiple labels depending on its local features. The value of each histogram bin of FC is a histogram of various features assigned to points in the bin region. We also introduce an efficient coding method to encode the local image features, call Radial Basis Coding (RBC). Combining RBC and FC together, and using a linear SVM classifier, our method is suitable for both image classification and object detection.

We propose a novel inference framework for finding maximal cliques in a weighted graph that satisfy hard constraints. The constraints specify the graph nodes that must belong to the solution as well as mutual exclusions of graph nodes, i.e., sets of nodes that cannot belong to the same solution. The proposed inference is based on a novel particle filter algorithm with state permeations. We apply the inference framework to a challenging problem of learning part-based, deformable object models. Two core problems in the learning framework, matching of image patches and finding salient parts, are formulated as two instances of the problem of finding maximal cliques with hard constraints. Our learning framework yields discriminative part based object models that achieve very good detection rate, and outperform other methods on object classes with large deformation.

We present a shape-based algorithm for detecting and recognizing non-rigid objects from natural images. The existing literature in this domain often cannot model the objects very well. In this paper, we use the skeleton (medial axis) information to capture the main structure of an object, which has the particular advantage in modeling articulation and non-rigid deformation. Given a set of training samples, a tree-union structure is learned on the extracted skeletons to model the variation in configuration. Each branch on the skeleton is associated with a few part-based templates, modeling the object boundary information. We then apply sum-and-max algorithm to perform rapid object detection by matching the skeleton-based active template to the edge map extracted from a test image. The algorithm reports the detection result by a composition of the local maximum responses. Compared with the alternatives on this topic, our algorithm requires less training samples. It is simple, yet efficient and effective. We show encouraging results on two widely used benchmark image sets: the Weizmann horse dataset [7] and the ETHZ dataset [16].