Zhe Chen

Wenhai Wang, Jifeng Dai, Zhe Chen et al.

Compared to the great progress of large-scale vision transformers (ViTs) in recent years, large-scale models based on convolutional neural networks (CNNs) are still in an early state. This work presents a new large-scale CNN-based foundation model, termed InternImage, which can obtain the gain from increasing parameters and training data like ViTs. Different from the recent CNNs that focus on large dense kernels, InternImage takes deformable convolution as the core operator, so that our model not only has the large effective receptive field required for downstream tasks such as detection and segmentation, but also has the adaptive spatial aggregation conditioned by input and task information. As a result, the proposed InternImage reduces the strict inductive bias of traditional CNNs and makes it possible to learn stronger and more robust patterns with large-scale parameters from massive data like ViTs. The effectiveness of our model is proven on challenging benchmarks including ImageNet, COCO, and ADE20K. It is worth mentioning that InternImage-H achieved a new record 65.4 mAP on COCO test-dev and 62.9 mIoU on ADE20K, outperforming current leading CNNs and ViTs. The code will be released at https://github.com/OpenGVLab/InternImage.

Zhe Chen, Yuchen Duan, Wenhai Wang et al.

This work investigates a simple yet powerful dense prediction task adapter for Vision Transformer (ViT). Unlike recently advanced variants that incorporate vision-specific inductive biases into their architectures, the plain ViT suffers inferior performance on dense predictions due to weak prior assumptions. To address this issue, we propose the ViT-Adapter, which allows plain ViT to achieve comparable performance to vision-specific transformers. Specifically, the backbone in our framework is a plain ViT that can learn powerful representations from large-scale multi-modal data. When transferring to downstream tasks, a pre-training-free adapter is used to introduce the image-related inductive biases into the model, making it suitable for these tasks. We verify ViT-Adapter on multiple dense prediction tasks, including object detection, instance segmentation, and semantic segmentation. Notably, without using extra detection data, our ViT-Adapter-L yields state-of-the-art 60.9 box AP and 53.0 mask AP on COCO test-dev. We hope that the ViT-Adapter could serve as an alternative for vision-specific transformers and facilitate future research. The code and models will be released at https://github.com/czczup/ViT-Adapter.

Liyao Tang, Yibing Zhan, Zhe Chen et al.

Point cloud segmentation is fundamental in understanding 3D environments. However, current 3D point cloud segmentation methods usually perform poorly on scene boundaries, which degenerates the overall segmentation performance. In this paper, we focus on the segmentation of scene boundaries. Accordingly, we first explore metrics to evaluate the segmentation performance on scene boundaries. To address the unsatisfactory performance on boundaries, we then propose a novel contrastive boundary learning (CBL) framework for point cloud segmentation. Specifically, the proposed CBL enhances feature discrimination between points across boundaries by contrasting their representations with the assistance of scene contexts at multiple scales. By applying CBL on three different baseline methods, we experimentally show that CBL consistently improves different baselines and assists them to achieve compelling performance on boundaries, as well as the overall performance, eg in mIoU. The experimental results demonstrate the effectiveness of our method and the importance of boundaries for 3D point cloud segmentation. Code and model will be made publicly available at https://github.com/LiyaoTang/contrastBoundary.

Weiyun Wang, Min Shi, Qingyun Li et al.

We present the All-Seeing (AS) project: a large-scale data and model for recognizing and understanding everything in the open world. Using a scalable data engine that incorporates human feedback and efficient models in the loop, we create a new dataset (AS-1B) with over 1 billion regions annotated with semantic tags, question-answering pairs, and detailed captions. It covers a wide range of 3.5 million common and rare concepts in the real world, and has 132.2 billion tokens that describe the concepts and their attributes. Leveraging this new dataset, we develop the All-Seeing model (ASM), a unified framework for panoptic visual recognition and understanding. The model is trained with open-ended language prompts and locations, which allows it to generalize to various vision and language tasks with remarkable zero-shot performance, including region-text retrieval, region recognition, captioning, and question-answering. We hope that this project can serve as a foundation for vision-language artificial general intelligence research. Models and the dataset shall be released at https://github.com/OpenGVLab/All-Seeing, and demo can be seen at https://huggingface.co/spaces/OpenGVLab/all-seeing.

Yuanyuan Liu, Wenbin Wang, Yibing Zhan et al.

Self-supervised facial representation has recently attracted increasing attention due to its ability to perform face understanding without relying on large-scale annotated datasets heavily. However, analytically, current contrastive-based self-supervised learning (SSL) still performs unsatisfactorily for learning facial representation. More specifically, existing contrastive learning (CL) tends to learn pose-invariant features that cannot depict the pose details of faces, compromising the learning performance. To conquer the above limitation of CL, we propose a novel Pose-disentangled Contrastive Learning (PCL) method for general self-supervised facial representation. Our PCL first devises a pose-disentangled decoder (PDD) with a delicately designed orthogonalizing regulation, which disentangles the pose-related features from the face-aware features; therefore, pose-related and other pose-unrelated facial information could be performed in individual subnetworks and do not affect each other's training. Furthermore, we introduce a pose-related contrastive learning scheme that learns pose-related information based on data augmentation of the same image, which would deliver more effective face-aware representation for various downstream tasks. We conducted linear evaluation on four challenging downstream facial understanding tasks, ie, facial expression recognition, face recognition, AU detection and head pose estimation. Experimental results demonstrate that our method significantly outperforms state-of-the-art SSL methods. Code is available at https://github.com/DreamMr/PCL}{https://github.com/DreamMr/PCL

Guo Chen, Sen Xing, Zhe Chen et al.

In this report, we present our champion solutions to five tracks at Ego4D challenge. We leverage our developed InternVideo, a video foundation model, for five Ego4D tasks, including Moment Queries, Natural Language Queries, Future Hand Prediction, State Change Object Detection, and Short-term Object Interaction Anticipation. InternVideo-Ego4D is an effective paradigm to adapt the strong foundation model to the downstream ego-centric video understanding tasks with simple head designs. In these five tasks, the performance of InternVideo-Ego4D comprehensively surpasses the baseline methods and the champions of CVPR2022, demonstrating the powerful representation ability of InternVideo as a video foundation model. Our code will be released at https://github.com/OpenGVLab/ego4d-eccv2022-solutions

Yuanfeng Ji, Zhe Chen, Enze Xie et al.

We propose a simple, efficient, yet powerful framework for dense visual predictions based on the conditional diffusion pipeline. Our approach follows a "noise-to-map" generative paradigm for prediction by progressively removing noise from a random Gaussian distribution, guided by the image. The method, called DDP, efficiently extends the denoising diffusion process into the modern perception pipeline. Without task-specific design and architecture customization, DDP is easy to generalize to most dense prediction tasks, e.g., semantic segmentation and depth estimation. In addition, DDP shows attractive properties such as dynamic inference and uncertainty awareness, in contrast to previous single-step discriminative methods. We show top results on three representative tasks with six diverse benchmarks, without tricks, DDP achieves state-of-the-art or competitive performance on each task compared to the specialist counterparts. For example, semantic segmentation (83.9 mIoU on Cityscapes), BEV map segmentation (70.6 mIoU on nuScenes), and depth estimation (0.05 REL on KITTI). We hope that our approach will serve as a solid baseline and facilitate future research

Shengyi Gao, Zhe Chen, Guo Chen et al.

In this report, we present our champion solution to the WSDM2023 Toloka Visual Question Answering (VQA) Challenge. Different from the common VQA and visual grounding (VG) tasks, this challenge involves a more complex scenario, i.e. inferring and locating the object implicitly specified by the given interrogative question. For this task, we leverage ViT-Adapter, a pre-training-free adapter network, to adapt multi-modal pre-trained Uni-Perceiver for better cross-modal localization. Our method ranks first on the leaderboard, achieving 77.5 and 76.347 IoU on public and private test sets, respectively. It shows that ViT-Adapter is also an effective paradigm for adapting the unified perception model to vision-language downstream tasks. Code and models will be released at https://github.com/czczup/ViT-Adapter/tree/main/wsdm2023.

Kai Chen, Enze Xie, Zhe Chen et al.

Diffusion models have attracted significant attention due to the remarkable ability to create content and generate data for tasks like image classification. However, the usage of diffusion models to generate the high-quality object detection data remains an underexplored area, where not only image-level perceptual quality but also geometric conditions such as bounding boxes and camera views are essential. Previous studies have utilized either copy-paste synthesis or layout-to-image (L2I) generation with specifically designed modules to encode the semantic layouts. In this paper, we propose the GeoDiffusion, a simple framework that can flexibly translate various geometric conditions into text prompts and empower pre-trained text-to-image (T2I) diffusion models for high-quality detection data generation. Unlike previous L2I methods, our GeoDiffusion is able to encode not only the bounding boxes but also extra geometric conditions such as camera views in self-driving scenes. Extensive experiments demonstrate GeoDiffusion outperforms previous L2I methods while maintaining 4x training time faster. To the best of our knowledge, this is the first work to adopt diffusion models for layout-to-image generation with geometric conditions and demonstrate that L2I-generated images can be beneficial for improving the performance of object detectors.

Xu Zhang, Wen Wang, Zhe Chen et al.

Animal pose estimation is challenging for existing image-based methods because of limited training data and large intra- and inter-species variances. Motivated by the progress of visual-language research, we propose that pre-trained language models (e.g., CLIP) can facilitate animal pose estimation by providing rich prior knowledge for describing animal keypoints in text. However, we found that building effective connections between pre-trained language models and visual animal keypoints is non-trivial since the gap between text-based descriptions and keypoint-based visual features about animal pose can be significant. To address this issue, we introduce a novel prompt-based Contrastive learning scheme for connecting Language and AniMal Pose (CLAMP) effectively. The CLAMP attempts to bridge the gap by adapting the text prompts to the animal keypoints during network training. The adaptation is decomposed into spatial-aware and feature-aware processes, and two novel contrastive losses are devised correspondingly. In practice, the CLAMP enables the first cross-modal animal pose estimation paradigm. Experimental results show that our method achieves state-of-the-art performance under the supervised, few-shot, and zero-shot settings, outperforming image-based methods by a large margin.

Xiao Wang, Zhe Chen, Bo Jiang et al.

To track the target in a video, current visual trackers usually adopt greedy search for target object localization in each frame, that is, the candidate region with the maximum response score will be selected as the tracking result of each frame. However, we found that this may be not an optimal choice, especially when encountering challenging tracking scenarios such as heavy occlusion and fast motion. To address this issue, we propose to maintain multiple tracking trajectories and apply beam search strategy for visual tracking, so that the trajectory with fewer accumulated errors can be identified. Accordingly, this paper introduces a novel multi-agent reinforcement learning based beam search tracking strategy, termed BeamTracking. It is mainly inspired by the image captioning task, which takes an image as input and generates diverse descriptions using beam search algorithm. Accordingly, we formulate the tracking as a sample selection problem fulfilled by multiple parallel decision-making processes, each of which aims at picking out one sample as their tracking result in each frame. Each maintained trajectory is associated with an agent to perform the decision-making and determine what actions should be taken to update related information. When all the frames are processed, we select the trajectory with the maximum accumulated score as the tracking result. Extensive experiments on seven popular tracking benchmark datasets validated the effectiveness of the proposed algorithm.

Zhe Chen, Jing Zhang, Yufei Xu et al.

Current object detectors typically have a feature pyramid (FP) module for multi-level feature fusion (MFF) which aims to mitigate the gap between features from different levels and form a comprehensive object representation to achieve better detection performance. However, they usually require heavy cross-level connections or iterative refinement to obtain better MFF result, making them complicated in structure and inefficient in computation. To address these issues, we propose a novel and efficient context modeling mechanism that can help existing FPs deliver better MFF results while reducing the computational costs effectively. In particular, we introduce a novel insight that comprehensive contexts can be decomposed and condensed into two types of representations for higher efficiency. The two representations include a locally concentrated representation and a globally summarized representation, where the former focuses on extracting context cues from nearby areas while the latter extracts key representations of the whole image scene as global context cues. By collecting the condensed contexts, we employ a Transformer decoder to investigate the relations between them and each local feature from the FP and then refine the MFF results accordingly. As a result, we obtain a simple and light-weight Transformer-based Context Condensation (TCC) module, which can boost various FPs and lower their computational costs simultaneously. Extensive experimental results on the challenging MS COCO dataset show that TCC is compatible to four representative FPs and consistently improves their detection accuracy by up to 7.8 % in terms of average precision and reduce their complexities by up to around 20% in terms of GFLOPs, helping them achieve state-of-the-art performance more efficiently. Code will be released.

Xiaoyi Dong, Pan Zhang, Yuhang Zang et al. · pku

The Large Vision-Language Model (LVLM) field has seen significant advancements, yet its progression has been hindered by challenges in comprehending fine-grained visual content due to limited resolution. Recent efforts have aimed to enhance the high-resolution understanding capabilities of LVLMs, yet they remain capped at approximately 1500 x 1500 pixels and constrained to a relatively narrow resolution range. This paper represents InternLM-XComposer2-4KHD, a groundbreaking exploration into elevating LVLM resolution capabilities up to 4K HD (3840 x 1600) and beyond. Concurrently, considering the ultra-high resolution may not be necessary in all scenarios, it supports a wide range of diverse resolutions from 336 pixels to 4K standard, significantly broadening its scope of applicability. Specifically, this research advances the patch division paradigm by introducing a novel extension: dynamic resolution with automatic patch configuration. It maintains the training image aspect ratios while automatically varying patch counts and configuring layouts based on a pre-trained Vision Transformer (ViT) (336 x 336), leading to dynamic training resolution from 336 pixels to 4K standard. Our research demonstrates that scaling training resolution up to 4K HD leads to consistent performance enhancements without hitting the ceiling of potential improvements. InternLM-XComposer2-4KHD shows superb capability that matches or even surpasses GPT-4V and Gemini Pro in 10 of the 16 benchmarks. The InternLM-XComposer2-4KHD model series with 7B parameters are publicly available at https://github.com/InternLM/InternLM-XComposer.

Zhe Chen, Daniel Harabor, Jiaoyang Li et al.

Multi-Agent Path Finding (MAPF) is a fundamental problem in robotics that asks us to compute collision-free paths for a team of agents, all moving across a shared map. Although many works appear on this topic, all current algorithms struggle as the number of agents grows. The principal reason is that existing approaches typically plan free-flow optimal paths, which creates congestion. To tackle this issue, we propose a new approach for MAPF where agents are guided to their destination by following congestion-avoiding paths. We evaluate the idea in two large-scale settings: one-shot MAPF, where each agent has a single destination, and lifelong MAPF, where agents are continuously assigned new destinations. Empirically, we report large improvements in solution quality for one-short MAPF and in overall throughput for lifelong MAPF.

Yuchen Duan, Weiyun Wang, Zhe Chen et al.

Transformers have revolutionized computer vision and natural language processing, but their high computational complexity limits their application in high-resolution image processing and long-context analysis. This paper introduces Vision-RWKV (VRWKV), a model adapted from the RWKV model used in the NLP field with necessary modifications for vision tasks. Similar to the Vision Transformer (ViT), our model is designed to efficiently handle sparse inputs and demonstrate robust global processing capabilities, while also scaling up effectively, accommodating both large-scale parameters and extensive datasets. Its distinctive advantage lies in its reduced spatial aggregation complexity, which renders it exceptionally adept at processing high-resolution images seamlessly, eliminating the necessity for windowing operations. Our evaluations demonstrate that VRWKV surpasses ViT's performance in image classification and has significantly faster speeds and lower memory usage processing high-resolution inputs. In dense prediction tasks, it outperforms window-based models, maintaining comparable speeds. These results highlight VRWKV's potential as a more efficient alternative for visual perception tasks. Code is released at https://github.com/OpenGVLab/Vision-RWKV.

Zhangwei Gao, Zhe Chen, Erfei Cui et al.

Multimodal large language models (MLLMs) have demonstrated impressive performance in vision-language tasks across a broad spectrum of domains. However, the large model scale and associated high computational costs pose significant challenges for training and deploying MLLMs on consumer-grade GPUs or edge devices, thereby hindering their widespread application. In this work, we introduce Mini-InternVL, a series of MLLMs with parameters ranging from 1B to 4B, which achieves 90% of the performance with only 5% of the parameters. This significant improvement in efficiency and effectiveness makes our models more accessible and applicable in various real-world scenarios. To further promote the adoption of our models, we develop a unified adaptation framework for Mini-InternVL, which enables our models to transfer and outperform specialized models in downstream tasks, including autonomous driving, medical images, and remote sensing. We believe that our study can provide valuable insights and resources to advance the development of efficient and effective MLLMs. Code is available at https://github.com/OpenGVLab/InternVL.

Hongcheng Liu, Zhe Chen, Hui Li et al.

Generating dialogue grounded in videos requires a high level of understanding and reasoning about the visual scenes in the videos. However, existing large visual-language models are not effective due to their latent features and decoder-only structure, especially with respect to spatio-temporal relationship reasoning. In this paper, we propose a novel approach named MSG-BART, which enhances the integration of video information by incorporating a multi-granularity spatio-temporal scene graph into an encoder-decoder pre-trained language model. Specifically, we integrate the global and local scene graph into the encoder and decoder, respectively, to improve both overall perception and target reasoning capability. To further improve the information selection capability, we propose a multi-pointer network to facilitate selection between text and video. Extensive experiments are conducted on three video-grounded dialogue benchmarks, which show the significant superiority of the proposed MSG-BART compared to a range of state-of-the-art approaches.

Ailing Yu, Lan Yao, Jingnan Liu et al.

Recent developments in Large Language Model (LLM)-based agents have shown impressive capabilities spanning multiple domains, exemplified by deep research systems that demonstrate superior performance on complex information-seeking and synthesis tasks. While general-purpose deep research agents have shown impressive capabilities, they struggle significantly with medical domain challenges, as evidenced by leading proprietary systems achieving limited accuracy on complex medical benchmarks. The key limitations are: (1) the model lacks sufficient dense medical knowledge for clinical reasoning, and (2) the framework is constrained by the absence of specialized retrieval tools tailored for medical contexts. We present a medical deep research agent that addresses these challenges through two core innovations. First, we develop a novel data synthesis framework using medical knowledge graphs, extracting the longest chains from subgraphs around rare medical entities to generate complex multi-hop question-answer pairs. Second, we integrate a custom-built private medical retrieval engine alongside general-purpose tools, enabling accurate medical information synthesis. Our approach generates 2100+ diverse trajectories across 12 medical specialties, each averaging 4.2 tool interactions. Through a two-stage training paradigm combining supervised fine-tuning and online reinforcement learning with composite rewards, our MedResearcher-R1-32B model demonstrates exceptional performance, establishing new state-of-the-art results on medical benchmarks while maintaining competitive performance on general deep research tasks. Our work demonstrates that strategic domain-specific innovations in architecture, tool design, and training data construction can enable smaller open-source models to outperform much larger proprietary systems in specialized domains.

Zhe Chen, Heyang Liu, Wenyi Yu et al.

Publishing open-source academic video recordings is an emergent and prevalent approach to sharing knowledge online. Such videos carry rich multimodal information including speech, the facial and body movements of the speakers, as well as the texts and pictures in the slides and possibly even the papers. Although multiple academic video datasets have been constructed and released, few of them support both multimodal content recognition and understanding tasks, which is partially due to the lack of high-quality human annotations. In this paper, we propose a novel multimodal, multigenre, and multipurpose audio-visual academic lecture dataset (M$^3$AV), which has almost 367 hours of videos from five sources covering computer science, mathematics, and medical and biology topics. With high-quality human annotations of the slide text and spoken words, in particular high-valued name entities, the dataset can be used for multiple audio-visual recognition and understanding tasks. Evaluations performed on contextual speech recognition, speech synthesis, and slide and script generation tasks demonstrate that the diversity of M$^3$AV makes it a challenging dataset.

Changyao Tian, Xizhou Zhu, Yuwen Xiong et al.

Developing generative models for interleaved image-text data has both research and practical value. It requires models to understand the interleaved sequences and subsequently generate images and text. However, existing attempts are limited by the issue that the fixed number of visual tokens cannot efficiently capture image details, which is particularly problematic in the multi-image scenarios. To address this, this paper presents MM-Interleaved, an end-to-end generative model for interleaved image-text data. It introduces a multi-scale and multi-image feature synchronizer module, allowing direct access to fine-grained image features in the previous context during the generation process. MM-Interleaved is end-to-end pre-trained on both paired and interleaved image-text corpora. It is further enhanced through a supervised fine-tuning phase, wherein the model improves its ability to follow complex multi-modal instructions. Experiments demonstrate the versatility of MM-Interleaved in recognizing visual details following multi-modal instructions and generating consistent images following both textual and visual conditions. Code and models are available at \url{https://github.com/OpenGVLab/MM-Interleaved}.

Liyao Tang, Zhe Chen, Shanshan Zhao et al.

Pseudo-labels are widely employed in weakly supervised 3D segmentation tasks where only sparse ground-truth labels are available for learning. Existing methods often rely on empirical label selection strategies, such as confidence thresholding, to generate beneficial pseudo-labels for model training. This approach may, however, hinder the comprehensive exploitation of unlabeled data points. We hypothesize that this selective usage arises from the noise in pseudo-labels generated on unlabeled data. The noise in pseudo-labels may result in significant discrepancies between pseudo-labels and model predictions, thus confusing and affecting the model training greatly. To address this issue, we propose a novel learning strategy to regularize the generated pseudo-labels and effectively narrow the gaps between pseudo-labels and model predictions. More specifically, our method introduces an Entropy Regularization loss and a Distribution Alignment loss for weakly supervised learning in 3D segmentation tasks, resulting in an ERDA learning strategy. Interestingly, by using KL distance to formulate the distribution alignment loss, it reduces to a deceptively simple cross-entropy-based loss which optimizes both the pseudo-label generation network and the 3D segmentation network simultaneously. Despite the simplicity, our method promisingly improves the performance. We validate the effectiveness through extensive experiments on various baselines and large-scale datasets. Results show that ERDA effectively enables the effective usage of all unlabeled data points for learning and achieves state-of-the-art performance under different settings. Remarkably, our method can outperform fully-supervised baselines using only 1% of true annotations. Code and model will be made publicly available at https://github.com/LiyaoTang/ERDA.

Wenhai Wang, Zhe Chen, Xiaokang Chen et al.

Large language models (LLMs) have notably accelerated progress towards artificial general intelligence (AGI), with their impressive zero-shot capacity for user-tailored tasks, endowing them with immense potential across a range of applications. However, in the field of computer vision, despite the availability of numerous powerful vision foundation models (VFMs), they are still restricted to tasks in a pre-defined form, struggling to match the open-ended task capabilities of LLMs. In this work, we present an LLM-based framework for vision-centric tasks, termed VisionLLM. This framework provides a unified perspective for vision and language tasks by treating images as a foreign language and aligning vision-centric tasks with language tasks that can be flexibly defined and managed using language instructions. An LLM-based decoder can then make appropriate predictions based on these instructions for open-ended tasks. Extensive experiments show that the proposed VisionLLM can achieve different levels of task customization through language instructions, from fine-grained object-level to coarse-grained task-level customization, all with good results. It's noteworthy that, with a generalist LLM-based framework, our model can achieve over 60\% mAP on COCO, on par with detection-specific models. We hope this model can set a new baseline for generalist vision and language models. The demo shall be released based on https://github.com/OpenGVLab/InternGPT. The code shall be released at https://github.com/OpenGVLab/VisionLLM.

Zhe Chen, Jing Zhang, Dacheng Tao

Although detection with Transformer (DETR) is increasingly popular, its global attention modeling requires an extremely long training period to optimize and achieve promising detection performance. Alternative to existing studies that mainly develop advanced feature or embedding designs to tackle the training issue, we point out that the Region-of-Interest (RoI) based detection refinement can easily help mitigate the difficulty of training for DETR methods. Based on this, we introduce a novel REcurrent Glimpse-based decOder (REGO) in this paper. In particular, the REGO employs a multi-stage recurrent processing structure to help the attention of DETR gradually focus on foreground objects more accurately. In each processing stage, visual features are extracted as glimpse features from RoIs with enlarged bounding box areas of detection results from the previous stage. Then, a glimpse-based decoder is introduced to provide refined detection results based on both the glimpse features and the attention modeling outputs of the previous stage. In practice, REGO can be easily embedded in representative DETR variants while maintaining their fully end-to-end training and inference pipelines. In particular, REGO helps Deformable DETR achieve 44.8 AP on the MSCOCO dataset with only 36 training epochs, compared with the first DETR and the Deformable DETR that require 500 and 50 epochs to achieve comparable performance, respectively. Experiments also show that REGO consistently boosts the performance of different DETR detectors by up to 7% relative gain at the same setting of 50 training epochs. Code is available via https://github.com/zhechen/Deformable-DETR-REGO.

Zhe Chen, Jiahao Wang, Wenhai Wang et al.

We propose an accurate and efficient scene text detection framework, termed FAST (i.e., faster arbitrarily-shaped text detector). Different from recent advanced text detectors that used complicated post-processing and hand-crafted network architectures, resulting in low inference speed, FAST has two new designs. (1) We design a minimalist kernel representation (only has 1-channel output) to model text with arbitrary shape, as well as a GPU-parallel post-processing to efficiently assemble text lines with a negligible time overhead. (2) We search the network architecture tailored for text detection, leading to more powerful features than most networks that are searched for image classification. Benefiting from these two designs, FAST achieves an excellent trade-off between accuracy and efficiency on several challenging datasets, including Total Text, CTW1500, ICDAR 2015, and MSRA-TD500. For example, FAST-T yields 81.6% F-measure at 152 FPS on Total-Text, outperforming the previous fastest method by 1.7 points and 70 FPS in terms of accuracy and speed. With TensorRT optimization, the inference speed can be further accelerated to over 600 FPS. Code and models will be released at https://github.com/czczup/FAST.

Xiao Wang, Jianing Li, Lin Zhu et al.

Different from visible cameras which record intensity images frame by frame, the biologically inspired event camera produces a stream of asynchronous and sparse events with much lower latency. In practice, visible cameras can better perceive texture details and slow motion, while event cameras can be free from motion blurs and have a larger dynamic range which enables them to work well under fast motion and low illumination. Therefore, the two sensors can cooperate with each other to achieve more reliable object tracking. In this work, we propose a large-scale Visible-Event benchmark (termed VisEvent) due to the lack of a realistic and scaled dataset for this task. Our dataset consists of 820 video pairs captured under low illumination, high speed, and background clutter scenarios, and it is divided into a training and a testing subset, each of which contains 500 and 320 videos, respectively. Based on VisEvent, we transform the event flows into event images and construct more than 30 baseline methods by extending current single-modality trackers into dual-modality versions. More importantly, we further build a simple but effective tracking algorithm by proposing a cross-modality transformer, to achieve more effective feature fusion between visible and event data. Extensive experiments on the proposed VisEvent dataset, FE108, COESOT, and two simulated datasets (i.e., OTB-DVS and VOT-DVS), validated the effectiveness of our model. The dataset and source code have been released on: \url{https://github.com/wangxiao5791509/VisEvent_SOT_Benchmark}.

Zhe Chen, Wenhai Wang, Enze Xie et al.

We present an extremely simple Ultra-Resolution Style Transfer framework, termed URST, to flexibly process arbitrary high-resolution images (e.g., 10000x10000 pixels) style transfer for the first time. Most of the existing state-of-the-art methods would fall short due to massive memory cost and small stroke size when processing ultra-high resolution images. URST completely avoids the memory problem caused by ultra-high resolution images by (1) dividing the image into small patches and (2) performing patch-wise style transfer with a novel Thumbnail Instance Normalization (TIN). Specifically, TIN can extract thumbnail features' normalization statistics and apply them to small patches, ensuring the style consistency among different patches. Overall, the URST framework has three merits compared to prior arts. (1) We divide input image into small patches and adopt TIN, successfully transferring image style with arbitrary high-resolution. (2) Experiments show that our URST surpasses existing SOTA methods on ultra-high resolution images benefiting from the effectiveness of the proposed stroke perceptual loss in enlarging the stroke size. (3) Our URST can be easily plugged into most existing style transfer methods and directly improve their performance even without training. Code is available at https://git.io/URST.

Zhe Chen, Jing Zhang, Dacheng Tao

Modern two-stage object detectors generally require excessively large models for their detection heads to achieve high accuracy. To address this problem, we propose that the model parameters of two-stage detection heads can be condensed and reduced by concentrating on object key parts. To this end, we first introduce an automatic object key part discovery task to make neural networks discover representative sub-parts in each foreground object. With these discovered key parts, we then decompose the object appearance modeling into a key part modeling process and a global modeling process for detection. Key part modeling encodes fine and detailed features from the discovered key parts, and global modeling encodes rough and holistic object characteristics. In practice, such decomposition allows us to significantly abridge model parameters without sacrificing much detection accuracy. Experiments on popular datasets illustrate that our proposed technique consistently maintains original performance while waiving around 50% of the model parameters of common two-stage detection heads, with the performance only deteriorating by 1.5% when waiving around 96% of the original model parameters. Codes are released on: https://github.com/zhechen/Condensing2stageDetection.

Zihan Fang, Zheng Lin, Zhe Chen et al.

Recently, there has been a surge in the development of advanced intelligent generative content (AIGC), especially large language models (LLMs). However, for many downstream tasks, it is necessary to fine-tune LLMs using private data. While federated learning offers a promising privacy-preserving solution to LLM fine-tuning, the substantial size of an LLM, combined with high computational and communication demands, makes it hard to apply to downstream tasks. More importantly, private edge servers often possess varying computing and network resources in real-world scenarios, introducing additional complexities to LLM fine-tuning. To tackle these problems, we design and implement an automated federated pipeline, named FedPipe, to fine-tune LLMs with minimal training cost but without adding any inference latency. FedPipe firstly identifies the weights to be fine-tuned based on their contributions to the LLM training. It then configures a low-rank adapter for each selected weight to train local low-rank adapters on an edge server, and aggregate local adapters of all edge servers to fine-tune the whole LLM. Finally, it appropriately quantizes the parameters of LLM to reduce memory space according to the requirements of edge servers. Extensive experiments demonstrate that FedPipe expedites the model training and achieves higher accuracy than state-of-the-art benchmarks.

Zhe Chen, Yusheng Liao, Shuyang Jiang et al.

Large language models hold promise for addressing medical challenges, such as medical diagnosis reasoning, research knowledge acquisition, clinical decision-making, and consumer health inquiry support. However, they often generate hallucinations due to limited medical knowledge. Incorporating external knowledge is therefore critical, which necessitates multi-source knowledge acquisition. We address this challenge by framing it as a source planning problem, which is to formulate context-appropriate queries tailored to the attributes of diverse sources. Existing approaches either overlook source planning or fail to achieve it effectively due to misalignment between the model's expectation of the sources and their actual content. To bridge this gap, we present MedOmniKB, a repository comprising multigenre and multi-structured medical knowledge sources. Leveraging these sources, we propose the Source Planning Optimisation method, which enhances multi-source utilisation. Our approach involves enabling an expert model to explore and evaluate potential plans while training a smaller model to learn source alignment. Experimental results demonstrate that our method substantially improves multi-source planning performance, enabling the optimised small model to achieve state-of-the-art results in leveraging diverse medical knowledge sources.

Tianbin Li, Yanzhou Su, Wei Li et al.

Despite significant advancements in general AI, its effectiveness in the medical domain is limited by the lack of specialized medical knowledge. To address this, we formulate GMAI-VL-5.5M, a multimodal medical dataset created by converting hundreds of specialized medical datasets with various annotations into high-quality image-text pairs. This dataset offers comprehensive task coverage, diverse modalities, and rich image-text data. Building upon this dataset, we develop GMAI-VL, a general medical vision-language model, with a three-stage training strategy that enhances the integration of visual and textual information. This approach significantly improves the model's ability to process multimodal data, supporting accurate diagnoses and clinical decision-making. Experiments show that GMAI-VL achieves state-of-the-art performance across various multimodal medical tasks, including visual question answering and medical image diagnosis.

Zhengren Wang, Jiayang Yu, Dongsheng Ma et al.

Domain-specific intelligence demands specialized knowledge and sophisticated reasoning for problem-solving, posing significant challenges for large language models (LLMs) that struggle with knowledge hallucination and inadequate reasoning capabilities under constrained parameter budgets. Inspired by Bloom's Taxonomy in educational theory, we propose Retrieval-Augmented Reasoning Modeling (RARE), a novel paradigm that decouples knowledge storage from reasoning optimization. RARE externalizes domain knowledge to retrievable sources and internalizes domain-specific reasoning patterns during training. Specifically, by injecting retrieved knowledge into training prompts with masked losses, RARE transforms learning objectives from rote memorization to contextualized reasoning. It enables models to bypass parameter-intensive memorization and prioritize the development of higher-order cognitive processes. Extensive experiments demonstrate that lightweight RARE-trained models (e.g., Llama-3.1-8B) could achieve state-of-the-art performance, surpassing retrieval-augmented GPT-4 and DeepSeek-R1 up to approximately 20\% accuracy. RARE establishes a paradigm shift where maintainable external knowledge bases synergize with compact, reasoning-optimized models, collectively driving more scalable domain-specific intelligence.

Bo Fu, Zhe Chen, Rahul Chandan et al.

We introduce the Block Rearrangement Problem (BRaP), a challenging component of large warehouse management which involves rearranging storage blocks within dense grids to achieve a target state. We formally define the BRaP as a graph search problem. Building on intuitions from sliding puzzle problems, we propose five search-based solution algorithms, leveraging joint configuration space search, classical planning, multi-agent pathfinding, and expert heuristics. We evaluate the five approaches empirically for plan quality and scalability. Despite the exponential relation between search space size and block number, our methods demonstrate efficiency in creating rearrangement plans for deeply buried blocks in up to 80x80 grids.

Zhe Chen, Yusheng Liao, Shuyang Jiang et al.

Medical large vision-language Models (Med-LVLMs) have shown promise in clinical applications but suffer from factual inaccuracies and unreliable outputs, posing risks in real-world diagnostics. While retrieval-augmented generation has emerged as a potential solution, current medical multimodal RAG systems are unable to perform effective retrieval across heterogeneous sources. The irrelevance of retrieved reports affects the factuality of analysis, while insufficient knowledge affects the credibility of clinical decision-making. To bridge the gap, we construct MedAtlas, which includes extensive multimodal report repositories and diverse text corpora. Based on it, we present HeteroRAG, a novel framework that enhances Med-LVLMs through heterogeneous knowledge sources. The framework introduces Modality-specific CLIPs for effective report retrieval and a Multi-corpora Query Generator for dynamically constructing queries for diverse corpora. Incorporating knowledge from such multifaceted sources, Med-LVLM is then trained with Heterogeneous Knowledge Preference Tuning to achieve cross-modality and multi-source knowledge alignment. Extensive experiments across 12 datasets and 3 modalities demonstrate that the proposed HeteroRAG achieves state-of-the-art performance in most medical vision language benchmarks, significantly improving factual accuracy and reliability of Med-LVLMs.

Bojie Shen, Zhe Chen, Muhammad Aamir Cheema et al.

Multi-Agent Path Finding (MAPF) is an important core problem for many new and emerging industrial applications. Many works appear on this topic each year, and a large number of substantial advancements and performance improvements have been reported. Yet measuring overall progress in MAPF is difficult: there are many potential competitors, and the computational burden for comprehensive experimentation is prohibitively large. Moreover, detailed data from past experimentation is usually unavailable. In this work, we introduce a set of methodological and visualisation tools which can help the community establish clear indicators for state-of-the-art MAPF performance and which can facilitate large-scale comparisons between MAPF solvers. Our objectives are to lower the barrier of entry for new researchers and to further promote the study of MAPF, since progress in the area and the main challenges are made much clearer.

Chen Chen, Zhe Chen, Jing Zhang et al.

Although point-based networks are demonstrated to be accurate for 3D point cloud modeling, they are still falling behind their voxel-based competitors in 3D detection. We observe that the prevailing set abstraction design for down-sampling points may maintain too much unimportant background information that can affect feature learning for detecting objects. To tackle this issue, we propose a novel set abstraction method named Semantics-Augmented Set Abstraction (SASA). Technically, we first add a binary segmentation module as the side output to help identify foreground points. Based on the estimated point-wise foreground scores, we then propose a semantics-guided point sampling algorithm to help retain more important foreground points during down-sampling. In practice, SASA shows to be effective in identifying valuable points related to foreground objects and improving feature learning for point-based 3D detection. Additionally, it is an easy-to-plug-in module and able to boost various point-based detectors, including single-stage and two-stage ones. Extensive experiments on the popular KITTI and nuScenes datasets validate the superiority of SASA, lifting point-based detection models to reach comparable performance to state-of-the-art voxel-based methods.

Xiao Wang, Zhe Chen, Jin Tang et al.

Most of the existing single object trackers track the target in a unitary local search window, making them particularly vulnerable to challenging factors such as heavy occlusions and out-of-view movements. Despite the attempts to further incorporate global search, prevailing mechanisms that cooperate local and global search are relatively static, thus are still sub-optimal for improving tracking performance. By further studying the local and global search results, we raise a question: can we allow more dynamics for cooperating both results? In this paper, we propose to introduce more dynamics by devising a dynamic attention-guided multi-trajectory tracking strategy. In particular, we construct dynamic appearance model that contains multiple target templates, each of which provides its own attention for locating the target in the new frame. Guided by different attention, we maintain diversified tracking results for the target to build multi-trajectory tracking history, allowing more candidates to represent the true target trajectory. After spanning the whole sequence, we introduce a multi-trajectory selection network to find the best trajectory that delivers improved tracking performance. Extensive experimental results show that our proposed tracking strategy achieves compelling performance on various large-scale tracking benchmarks. The project page of this paper can be found at https://sites.google.com/view/mt-track/.

Jack Humphreys, Zhe Chen, Dacheng Tao

Action recognition, which is formulated as a task to identify various human actions in a video, has attracted increasing interest from computer vision researchers due to its importance in various applications. Recently, appearance-based methods have achieved promising progress towards accurate action recognition. In general, these methods mainly fulfill the task by applying various schemes to model spatial and temporal visual information effectively. To better understand the current progress of appearance-based action recognition, we provide a comprehensive review of recent achievements in this area. In particular, we summarise and discuss several dozens of related research papers, which can be roughly divided into four categories according to different appearance modelling strategies. The obtained categories include 2D convolutional methods, 3D convolutional methods, motion representation-based methods, and context representation-based methods. We analyse and discuss representative methods from each category, comprehensively. Empirical results are also summarised to better illustrate cutting-edge algorithms. We conclude by identifying important areas for future research gleaned from our categorisation.

Jing Zhang, Zhe Chen, Dacheng Tao

Human keypoint detection from a single image is very challenging due to occlusion, blur, illumination and scale variance. In this paper, we address this problem from three aspects by devising an efficient network structure, proposing three effective training strategies, and exploiting four useful postprocessing techniques. First, we find that context information plays an important role in reasoning human body configuration and invisible keypoints. Inspired by this, we propose a cascaded context mixer (CCM), which efficiently integrates spatial and channel context information and progressively refines them. Then, to maximize CCM's representation capability, we develop a hard-negative person detection mining strategy and a joint-training strategy by exploiting abundant unlabeled data. It enables CCM to learn discriminative features from massive diverse poses. Third, we present several sub-pixel refinement techniques for postprocessing keypoint predictions to improve detection accuracy. Extensive experiments on the MS COCO keypoint detection benchmark demonstrate the superiority of the proposed method over representative state-of-the-art (SOTA) methods. Our single model achieves comparable performance with the winner of the 2018 COCO Keypoint Detection Challenge. The final ensemble model sets a new SOTA on this benchmark.

Jing Zhang, Zhe Chen, Dacheng Tao

Human keypoint detection from a single image is very challenging due to occlusion, blur, illumination and scale variance of person instances. In this paper, we find that context information plays an important role in addressing these issues, and propose a novel method named progressive context refinement (PCR) for human keypoint detection. First, we devise a simple but effective context-aware module (CAM) that can efficiently integrate spatial and channel context information to aid feature learning for locating hard keypoints. Then, we construct the PCR model by stacking several CAMs sequentially with shortcuts and employ multi-task learning to progressively refine the context information and predictions. Besides, to maximize PCR's potential for the aforementioned hard case inference, we propose a hard-negative person detection mining strategy together with a joint-training strategy by exploiting the unlabeled coco dataset and external dataset. Extensive experiments on the COCO keypoint detection benchmark demonstrate the superiority of PCR over representative state-of-the-art (SOTA) methods. Our single model achieves comparable performance with the winner of the 2018 COCO Keypoint Detection Challenge. The final ensemble model sets a new SOTA on this benchmark.

Zhe Chen, Jing Zhang, Dacheng Tao

Despite rapid developments in visual image-based road detection, robustly identifying road areas in visual images remains challenging due to issues like illumination changes and blurry images. To this end, LiDAR sensor data can be incorporated to improve the visual image-based road detection, because LiDAR data is less susceptible to visual noises. However, the main difficulty in introducing LiDAR information into visual image-based road detection is that LiDAR data and its extracted features do not share the same space with the visual data and visual features. Such gaps in spaces may limit the benefits of LiDAR information for road detection. To overcome this issue, we introduce a novel Progressive LiDAR Adaptation-aided Road Detection (PLARD) approach to adapt LiDAR information into visual image-based road detection and improve detection performance. In PLARD, progressive LiDAR adaptation consists of two subsequent modules: 1) data space adaptation, which transforms the LiDAR data to the visual data space to align with the perspective view by applying altitude difference-based transformation; and 2) feature space adaptation, which adapts LiDAR features to visual features through a cascaded fusion structure. Comprehensive empirical studies on the well-known KITTI road detection benchmark demonstrate that PLARD takes advantage of both the visual and LiDAR information, achieving much more robust road detection even in challenging urban scenes. In particular, PLARD outperforms other state-of-the-art road detection models and is currently top of the publicly accessible benchmark leader-board.

Zhe Chen, Zhibin Hong, Dacheng Tao

Over these years, Correlation Filter-based Trackers (CFTs) have aroused increasing interests in the field of visual object tracking, and have achieved extremely compelling results in different competitions and benchmarks. In this paper, our goal is to review the developments of CFTs with extensive experimental results. 11 trackers are surveyed in our work, based on which a general framework is summarized. Furthermore, we investigate different training schemes for correlation filters, and also discuss various effective improvements that have been made recently. Comprehensive experiments have been conducted to evaluate the effectiveness and efficiency of the surveyed CFTs, and comparisons have been made with other competing trackers. The experimental results have shown that state-of-art performance, in terms of robustness, speed and accuracy, can be achieved by several recent CFTs, such as MUSTer and SAMF. We find that further improvements for correlation filter-based tracking can be made on estimating scales, applying part-based tracking strategy and cooperating with long-term tracking methods.