Yujie Zhong

Chang Liu, Haoning Wu, Yujie Zhong et al.

Generative models have recently exhibited exceptional capabilities in text-to-image generation, but still struggle to generate image sequences coherently. In this work, we focus on a novel, yet challenging task of generating a coherent image sequence based on a given storyline, denoted as open-ended visual storytelling. We make the following three contributions: (i) to fulfill the task of visual storytelling, we propose a learning-based auto-regressive image generation model, termed as StoryGen, with a novel vision-language context module, that enables to generate the current frame by conditioning on the corresponding text prompt and preceding image-caption pairs; (ii) to address the data shortage of visual storytelling, we collect paired image-text sequences by sourcing from online videos and open-source E-books, establishing processing pipeline for constructing a large-scale dataset with diverse characters, storylines, and artistic styles, named StorySalon; (iii) Quantitative experiments and human evaluations have validated the superiority of our StoryGen, where we show StoryGen can generalize to unseen characters without any optimization, and generate image sequences with coherent content and consistent character. Code, dataset, and models are available at https://haoningwu3639.github.io/StoryGen_Webpage/

Dingfeng Shi, Yujie Zhong, Qiong Cao et al.

This work aims at advancing temporal action detection (TAD) using an encoder-decoder framework with action queries, similar to DETR, which has shown great success in object detection. However, the framework suffers from several problems if directly applied to TAD: the insufficient exploration of inter-query relation in the decoder, the inadequate classification training due to a limited number of training samples, and the unreliable classification scores at inference. To this end, we first propose a relational attention mechanism in the decoder, which guides the attention among queries based on their relations. Moreover, we propose two losses to facilitate and stabilize the training of action classification. Lastly, we propose to predict the localization quality of each action query at inference in order to distinguish high-quality queries. The proposed method, named ReAct, achieves the state-of-the-art performance on THUMOS14, with much lower computational costs than previous methods. Besides, extensive ablation studies are conducted to verify the effectiveness of each proposed component. The code is available at https://github.com/sssste/React.

Dingfeng Shi, Yujie Zhong, Qiong Cao et al.

In this paper, we present a one-stage framework TriDet for temporal action detection. Existing methods often suffer from imprecise boundary predictions due to the ambiguous action boundaries in videos. To alleviate this problem, we propose a novel Trident-head to model the action boundary via an estimated relative probability distribution around the boundary. In the feature pyramid of TriDet, we propose an efficient Scalable-Granularity Perception (SGP) layer to mitigate the rank loss problem of self-attention that takes place in the video features and aggregate information across different temporal granularities. Benefiting from the Trident-head and the SGP-based feature pyramid, TriDet achieves state-of-the-art performance on three challenging benchmarks: THUMOS14, HACS and EPIC-KITCHEN 100, with lower computational costs, compared to previous methods. For example, TriDet hits an average mAP of $69.3\%$ on THUMOS14, outperforming the previous best by $2.5\%$, but with only $74.6\%$ of its latency. The code is released to https://github.com/sssste/TriDet.

Cong Han, Yujie Zhong, Dengjie Li et al.

Recently, the open-vocabulary semantic segmentation problem has attracted increasing attention and the best performing methods are based on two-stream networks: one stream for proposal mask generation and the other for segment classification using a pretrained visual-language model. However, existing two-stream methods require passing a great number of (up to a hundred) image crops into the visual-language model, which is highly inefficient. To address the problem, we propose a network that only needs a single pass through the visual-language model for each input image. Specifically, we first propose a novel network adaptation approach, termed patch severance, to restrict the harmful interference between the patch embeddings in the pre-trained visual encoder. We then propose classification anchor learning to encourage the network to spatially focus on more discriminative features for classification. Extensive experiments demonstrate that the proposed method achieves outstanding performance, surpassing state-of-the-art methods while being 4 to 7 times faster at inference. Code: https://github.com/CongHan0808/DeOP.git

Sheng Guo, Zihua Xiong, Yujie Zhong et al.

In this paper, we present a new cross-architecture contrastive learning (CACL) framework for self-supervised video representation learning. CACL consists of a 3D CNN and a video transformer which are used in parallel to generate diverse positive pairs for contrastive learning. This allows the model to learn strong representations from such diverse yet meaningful pairs. Furthermore, we introduce a temporal self-supervised learning module able to predict an Edit distance explicitly between two video sequences in the temporal order. This enables the model to learn a rich temporal representation that compensates strongly to the video-level representation learned by the CACL. We evaluate our method on the tasks of video retrieval and action recognition on UCF101 and HMDB51 datasets, where our method achieves excellent performance, surpassing the state-of-the-art methods such as VideoMoCo and MoCo+BE by a large margin. The code is made available at https://github.com/guoshengcv/CACL.

Anthony Cioppa, Silvio Giancola, Vladimir Somers et al. · pku

The SoccerNet 2023 challenges were the third annual video understanding challenges organized by the SoccerNet team. For this third edition, the challenges were composed of seven vision-based tasks split into three main themes. The first theme, broadcast video understanding, is composed of three high-level tasks related to describing events occurring in the video broadcasts: (1) action spotting, focusing on retrieving all timestamps related to global actions in soccer, (2) ball action spotting, focusing on retrieving all timestamps related to the soccer ball change of state, and (3) dense video captioning, focusing on describing the broadcast with natural language and anchored timestamps. The second theme, field understanding, relates to the single task of (4) camera calibration, focusing on retrieving the intrinsic and extrinsic camera parameters from images. The third and last theme, player understanding, is composed of three low-level tasks related to extracting information about the players: (5) re-identification, focusing on retrieving the same players across multiple views, (6) multiple object tracking, focusing on tracking players and the ball through unedited video streams, and (7) jersey number recognition, focusing on recognizing the jersey number of players from tracklets. Compared to the previous editions of the SoccerNet challenges, tasks (2-3-7) are novel, including new annotations and data, task (4) was enhanced with more data and annotations, and task (6) now focuses on end-to-end approaches. More information on the tasks, challenges, and leaderboards are available on https://www.soccer-net.org. Baselines and development kits can be found on https://github.com/SoccerNet.

Xiao Zhou, Yujie Zhong, Zhen Cheng et al.

Object re-identification (ReID) aims to find instances with the same identity as the given probe from a large gallery. Pairwise losses play an important role in training a strong ReID network. Existing pairwise losses densely exploit each instance as an anchor and sample its triplets in a mini-batch. This dense sampling mechanism inevitably introduces positive pairs that share few visual similarities, which can be harmful to the training. To address this problem, we propose a novel loss paradigm termed Sparse Pairwise (SP) loss that only leverages few appropriate pairs for each class in a mini-batch, and empirically demonstrate that it is sufficient for the ReID tasks. Based on the proposed loss framework, we propose an adaptive positive mining strategy that can dynamically adapt to diverse intra-class variations. Extensive experiments show that SP loss and its adaptive variant AdaSP loss outperform other pairwise losses, and achieve state-of-the-art performance across several ReID benchmarks. Code is available at https://github.com/Astaxanthin/AdaSP.

Chang Liu, Yujie Zhong, Andrew Zisserman et al.

In this paper, we consider the problem of generalised visual object counting, with the goal of developing a computational model for counting the number of objects from arbitrary semantic categories, using arbitrary number of "exemplars", i.e. zero-shot or few-shot counting. To this end, we make the following four contributions: (1) We introduce a novel transformer-based architecture for generalised visual object counting, termed as Counting Transformer (CounTR), which explicitly capture the similarity between image patches or with given "exemplars" with the attention mechanism;(2) We adopt a two-stage training regime, that first pre-trains the model with self-supervised learning, and followed by supervised fine-tuning;(3) We propose a simple, scalable pipeline for synthesizing training images with a large number of instances or that from different semantic categories, explicitly forcing the model to make use of the given "exemplars";(4) We conduct thorough ablation studies on the large-scale counting benchmark, e.g. FSC-147, and demonstrate state-of-the-art performance on both zero and few-shot settings.

Silvio Giancola, Anthony Cioppa, Adrien Deliège et al.

The SoccerNet 2022 challenges were the second annual video understanding challenges organized by the SoccerNet team. In 2022, the challenges were composed of 6 vision-based tasks: (1) action spotting, focusing on retrieving action timestamps in long untrimmed videos, (2) replay grounding, focusing on retrieving the live moment of an action shown in a replay, (3) pitch localization, focusing on detecting line and goal part elements, (4) camera calibration, dedicated to retrieving the intrinsic and extrinsic camera parameters, (5) player re-identification, focusing on retrieving the same players across multiple views, and (6) multiple object tracking, focusing on tracking players and the ball through unedited video streams. Compared to last year's challenges, tasks (1-2) had their evaluation metrics redefined to consider tighter temporal accuracies, and tasks (3-6) were novel, including their underlying data and annotations. More information on the tasks, challenges and leaderboards are available on https://www.soccer-net.org. Baselines and development kits are available on https://github.com/SoccerNet.

Xianing Chen, Qiong Cao, Yujie Zhong et al.

Transformers are successfully applied to computer vision due to their powerful modeling capacity with self-attention. However, the excellent performance of transformers heavily depends on enormous training images. Thus, a data-efficient transformer solution is urgently needed. In this work, we propose an early knowledge distillation framework, which is termed as DearKD, to improve the data efficiency required by transformers. Our DearKD is a two-stage framework that first distills the inductive biases from the early intermediate layers of a CNN and then gives the transformer full play by training without distillation. Further, our DearKD can be readily applied to the extreme data-free case where no real images are available. In this case, we propose a boundary-preserving intra-divergence loss based on DeepInversion to further close the performance gap against the full-data counterpart. Extensive experiments on ImageNet, partial ImageNet, data-free setting and other downstream tasks prove the superiority of DearKD over its baselines and state-of-the-art methods.

Changcheng Xiao, Qiong Cao, Yujie Zhong et al.

Significant progress has been achieved in multi-object tracking (MOT) through the evolution of detection and re-identification (ReID) techniques. Despite these advancements, accurately tracking objects in scenarios with homogeneous appearance and heterogeneous motion remains a challenge. This challenge arises from two main factors: the insufficient discriminability of ReID features and the predominant utilization of linear motion models in MOT. In this context, we introduce a novel motion-based tracker, MotionTrack, centered around a learnable motion predictor that relies solely on object trajectory information. This predictor comprehensively integrates two levels of granularity in motion features to enhance the modeling of temporal dynamics and facilitate precise future motion prediction for individual objects. Specifically, the proposed approach adopts a self-attention mechanism to capture token-level information and a Dynamic MLP layer to model channel-level features. MotionTrack is a simple, online tracking approach. Our experimental results demonstrate that MotionTrack yields state-of-the-art performance on datasets such as Dancetrack and SportsMOT, characterized by highly complex object motion.

Chengjian Feng, Zequn Jie, Yujie Zhong et al.

Recent LSS-based multi-view 3D object detection has made tremendous progress, by processing the features in Brid-Eye-View (BEV) via the convolutional detector. However, the typical convolution ignores the radial symmetry of the BEV features and increases the difficulty of the detector optimization. To preserve the inherent property of the BEV features and ease the optimization, we propose an azimuth-equivariant convolution (AeConv) and an azimuth-equivariant anchor. The sampling grid of AeConv is always in the radial direction, thus it can learn azimuth-invariant BEV features. The proposed anchor enables the detection head to learn predicting azimuth-irrelevant targets. In addition, we introduce a camera-decoupled virtual depth to unify the depth prediction for the images with different camera intrinsic parameters. The resultant detector is dubbed Azimuth-equivariant Detector (AeDet). Extensive experiments are conducted on nuScenes, and AeDet achieves a 62.0% NDS, surpassing the recent multi-view 3D object detectors such as PETRv2 and BEVDepth by a large margin. Project page: https://fcjian.github.io/aedet.

Yiheng Lin, Siyu Jiao, Xiaohan Lan et al.

Recent advances in Multimodal Large Language Models (MLLMs) and diffusion-based generative models have substantially improved prompt-driven image editing. However, scene text editing remains challenging, as it requires models to precisely modify textual content while preserving visual realism and non-target regions. Current open-source models still lag behind proprietary systems, largely due to the scarcity of high-quality training data and the lack of standardized benchmarks tailored to text editing. To address these challenges, we present TextSculptor, a comprehensive framework for data construction and evaluation of scene text editing. We first develop an automated data construction pipeline that combines text-aware image synthesis with programmatic text rendering and compositing. Based on this pipeline, we build TextSculpt-Data, a large-scale dataset containing 3.2M training samples, including 1.2M OCR-verified text-to-image samples and 2M paired text editing samples with naturally aligned source-target images and strong background consistency. We further introduce TextSculpt-Bench, a benchmark covering four fundamental text editing tasks: text addition, text replacement, text removal, and hybrid editing. To support reliable evaluation, we design a tailored protocol that measures text accuracy, visual quality, and background preservation through OCR-based text alignment, multimodal judgment, and background-region similarity. Extensive experiments show that TextSculptor improves open-source text editing performance and narrows the gap to proprietary models. The data and benchmark are available at https://github.com/linyiheng123/TextSculptor.

Chang-Bin Zhang, Yujie Zhong, Qiang Zhang et al.

While visually grounded Chain-of-Thought (CoT) has emerged as a promising paradigm to enhance fine-grained perception in multimodal large language models (MLLMs), its efficacy during the inference phase remains underexplored. In this work, we empirically find that mandating explicit object boxes in visually grounded CoT during inference often degrades performance compared to standard textual CoT, which reasons without explicit visual grounding. We hypothesize that the visual localization capability can be internalized into the textual CoT and that the mandatory explicit grounding introduces unnecessary interference with the model's primary objective of answer prediction. To address this problem, we propose Internalizing Visually Grounded Reasoning (\textbf{iVGR}), a novel reinforcement learning framework that transfers localization capabilities into the textual reasoning process. We employ a dual-stream training strategy, where a textual stream is aligned with a high-quality visually grounded stream via a proposed consistency reward, enabling the model to localize accurately without explicit grounding during inference. Extensive experiments demonstrate that our method significantly outperforms existing baselines on fine-grained benchmarks, while maintaining the flexibility to support tool-assisted inference workflows.

Zixu Wang, Yujie Zhong, Yishu Miao et al.

Despite recent progress in video and language representation learning, the weak or sparse correspondence between the two modalities remains a bottleneck in the area. Most video-language models are trained via pair-level loss to predict whether a pair of video and text is aligned. However, even in paired video-text segments, only a subset of the frames are semantically relevant to the corresponding text, with the remainder representing noise; where the ratio of noisy frames is higher for longer videos. We propose FineCo (Fine-grained Contrastive Loss for Frame Sampling), an approach to better learn video and language representations with a fine-grained contrastive objective operating on video frames. It helps distil a video by selecting the frames that are semantically equivalent to the text, improving cross-modal correspondence. Building on the well established VideoCLIP model as a starting point, FineCo achieves state-of-the-art performance on YouCookII, a text-video retrieval benchmark with long videos. FineCo also achieves competitive results on text-video retrieval (MSR-VTT), and video question answering datasets (MSR-VTT QA and MSR-VTT MC) with shorter videos.

Siyu Jiao, Yiheng Lin, Yujie Zhong et al.

Recent progress in Multimodal Large Language Models (MLLMs) demonstrates that Chain-of-Thought (CoT) reasoning enables systematic solutions to complex understanding tasks. However, its extension to generation tasks remains nascent and limited by scenario-specific mechanisms that hinder generalization and adaptation. In this work, we present ThinkGen, the first think-driven visual generation framework that explicitly leverages MLLM's CoT reasoning in various generation scenarios. ThinkGen employs a decoupled architecture comprising a pretrained MLLM and a Diffusion Transformer (DiT), wherein the MLLM generates tailored instructions based on user intent, and DiT produces high-quality images guided by these instructions. We further propose a separable GRPO-based training paradigm (SepGRPO), alternating reinforcement learning between the MLLM and DiT modules. This flexible design enables joint training across diverse datasets, facilitating effective CoT reasoning for a wide range of generative scenarios. Extensive experiments demonstrate that ThinkGen achieves robust, state-of-the-art performance across multiple generation benchmarks. Code is available: https://github.com/jiaosiyuu/ThinkGen

Dengjie Li, Siyu Chen, Yujie Zhong et al.

In person re-identification (ReID) tasks, many works explore the learning of part features to improve the performance over global image features. Existing methods explicitly extract part features by either using a hand-designed image division or keypoints obtained with external visual systems. In this work, we propose to learn Discriminative implicit Parts (DiPs) which are decoupled from explicit body parts. Therefore, DiPs can learn to extract any discriminative features that can benefit in distinguishing identities, which is beyond predefined body parts (such as accessories). Moreover, we propose a novel implicit position to give a geometric interpretation for each DiP. The implicit position can also serve as a learning signal to encourage DiPs to be more position-equivariant with the identity in the image. Lastly, an additional DiP weighting is introduced to handle the invisible or occluded situation and further improve the feature representation of DiPs. Extensive experiments show that the proposed method achieves state-of-the-art performance on multiple person ReID benchmarks.

Lei Chen, Feng Yan, Yujie Zhong et al.

Multimodal Large Language Models (MLLM) have made significant progress in the field of document analysis. Despite this, existing benchmarks typically focus only on extracting text and simple layout information, neglecting the complex interactions between elements in structured documents such as mind maps and flowcharts. To address this issue, we introduce the new benchmark named MindBench, which not only includes meticulously constructed bilingual authentic or synthetic images, detailed annotations, evaluation metrics and baseline models, but also specifically designs five types of structured understanding and parsing tasks. These tasks include full parsing, partial parsing, position-related parsing, structured Visual Question Answering (VQA), and position-related VQA, covering key areas such as text recognition, spatial awareness, relationship discernment, and structured parsing. Extensive experimental results demonstrate the substantial potential and significant room for improvement in current models' ability to handle structured document information. We anticipate that the launch of MindBench will significantly advance research and application development in structured document analysis technology. MindBench is available at: https://miasanlei.github.io/MindBench.github.io/.

Dingfeng Shi, Qiong Cao, Yujie Zhong et al.

Temporal action detection (TAD) aims to detect all action boundaries and their corresponding categories in an untrimmed video. The unclear boundaries of actions in videos often result in imprecise predictions of action boundaries by existing methods. To resolve this issue, we propose a one-stage framework named TriDet. First, we propose a Trident-head to model the action boundary via an estimated relative probability distribution around the boundary. Then, we analyze the rank-loss problem (i.e. instant discriminability deterioration) in transformer-based methods and propose an efficient scalable-granularity perception (SGP) layer to mitigate this issue. To further push the limit of instant discriminability in the video backbone, we leverage the strong representation capability of pretrained large models and investigate their performance on TAD. Last, considering the adequate spatial-temporal context for classification, we design a decoupled feature pyramid network with separate feature pyramids to incorporate rich spatial context from the large model for localization. Experimental results demonstrate the robustness of TriDet and its state-of-the-art performance on multiple TAD datasets, including hierarchical (multilabel) TAD datasets.

Cong Wei, Haoxian Tan, Yujie Zhong et al.

Recent advancements have empowered Large Language Models for Vision (vLLMs) to generate detailed perceptual outcomes, including bounding boxes and masks. Nonetheless, there are two constraints that restrict the further application of these vLLMs: the incapability of handling multiple targets per query and the failure to identify the absence of query objects in the image. In this study, we acknowledge that the main cause of these problems is the insufficient complexity of training queries. Consequently, we define the general sequence format for complex queries. Then we incorporate a semantic segmentation task in the current pipeline to fulfill the requirements of training data. Furthermore, we present three novel strategies to effectively handle the challenges arising from the direct integration of the proposed format. The effectiveness of our model in processing complex queries is validated by the comparable results with conventional methods on both close-set and open-set semantic segmentation datasets. Additionally, we outperform a series of vLLMs in reasoning and referring segmentation, showcasing our model's remarkable capabilities. We release the code at https://github.com/congvvc/LaSagnA.

Cong Wei, Yujie Zhong, Haoxian Tan et al.

This paper aims to address universal segmentation for image and video perception with the strong reasoning ability empowered by Visual Large Language Models (VLLMs). Despite significant progress in current unified segmentation methods, limitations in adaptation to both image and video scenarios, as well as the complex reasoning segmentation, make it difficult for them to handle various challenging instructions and achieve an accurate understanding of fine-grained vision-language correlations. We propose HyperSeg, the first VLLM-based universal segmentation model for pixel-level image and video perception, encompassing generic segmentation tasks and more complex reasoning perception tasks requiring powerful reasoning abilities and world knowledge. Besides, to fully leverage the recognition capabilities of VLLMs and the fine-grained visual information, HyperSeg incorporates hybrid entity recognition and fine-grained visual perceiver modules for various segmentation tasks. Combined with the temporal adapter, HyperSeg achieves a comprehensive understanding of temporal information. Experimental results validate the effectiveness of our insights in resolving universal image and video segmentation tasks, including the more complex reasoning perception tasks. Our code is available.

Yingsen Zeng, Yujie Zhong, Chengjian Feng et al.

Temporal Action Detection (TAD) focuses on detecting pre-defined actions, while Moment Retrieval (MR) aims to identify the events described by open-ended natural language within untrimmed videos. Despite that they focus on different events, we observe they have a significant connection. For instance, most descriptions in MR involve multiple actions from TAD. In this paper, we aim to investigate the potential synergy between TAD and MR. Firstly, we propose a unified architecture, termed Unified Moment Detection (UniMD), for both TAD and MR. It transforms the inputs of the two tasks, namely actions for TAD or events for MR, into a common embedding space, and utilizes two novel query-dependent decoders to generate a uniform output of classification score and temporal segments. Secondly, we explore the efficacy of two task fusion learning approaches, pre-training and co-training, in order to enhance the mutual benefits between TAD and MR. Extensive experiments demonstrate that the proposed task fusion learning scheme enables the two tasks to help each other and outperform the separately trained counterparts. Impressively, UniMD achieves state-of-the-art results on three paired datasets Ego4D, Charades-STA, and ActivityNet. Our code is available at https://github.com/yingsen1/UniMD.

Cong Wei, Yujie Zhong, Haoxian Tan et al.

Boosted by Multi-modal Large Language Models (MLLMs), text-guided universal segmentation models for the image and video domains have made rapid progress recently. However, these methods are often developed separately for specific domains, overlooking the similarities in task settings and solutions across these two areas. In this paper, we define the union of referring segmentation and reasoning segmentation at both the image and video levels as Instructed Visual Segmentation (IVS). Correspondingly, we propose InstructSeg, an end-to-end segmentation pipeline equipped with MLLMs for IVS. Specifically, we employ an object-aware video perceiver to extract temporal and object information from reference frames, facilitating comprehensive video understanding. Additionally, we introduce vision-guided multi-granularity text fusion to better integrate global and detailed text information with fine-grained visual guidance. By leveraging multi-task and end-to-end training, InstructSeg demonstrates superior performance across diverse image and video segmentation tasks, surpassing both segmentation specialists and MLLM-based methods with a single model. Our code is available at https://github.com/congvvc/InstructSeg.

Zhijian Huang, Chengjian Feng, Feng Yan et al.

Large Multimodal Models (LMMs) have demonstrated exceptional comprehension and interpretation capabilities in Autonomous Driving (AD) by incorporating large language models. Despite the advancements, current data-driven AD approaches tend to concentrate on a single dataset and specific tasks, neglecting their overall capabilities and ability to generalize. To bridge these gaps, we propose RoboTron-Drive, a general large multimodal model designed to process diverse data inputs, such as images and multi-view videos, while performing a broad spectrum of AD tasks, including perception, prediction, and planning. Initially, the model undergoes curriculum pre-training to process varied visual signals and perform basic visual comprehension and perception tasks. Subsequently, we augment and standardize various AD datasets to finetune the model, resulting in an all-in-one LMM for autonomous driving. To assess the general capabilities and generalization ability, we conduct evaluations on six public benchmarks and undertake zero-shot transfer on three unseen datasets, where RoboTron-Drive achieves state-of-the-art performance across all tasks. We hope RoboTron-Drive as a promising solution for AD in the real world. Project page with code: https://github.com/zhijian11/RoboTron-Drive.

Jiancheng Huang, Yu Gao, Zequn Jie et al.

In this paper, we introduce MRStyle, a comprehensive framework that enables color style transfer using multi-modality reference, including image and text. To achieve a unified style feature space for both modalities, we first develop a neural network called IRStyle, which generates stylized 3D lookup tables for image reference. This is accomplished by integrating an interaction dual-mapping network with a combined supervised learning pipeline, resulting in three key benefits: elimination of visual artifacts, efficient handling of high-resolution images with low memory usage, and maintenance of style consistency even in situations with significant color style variations. For text reference, we align the text feature of stable diffusion priors with the style feature of our IRStyle to perform text-guided color style transfer (TRStyle). Our TRStyle method is highly efficient in both training and inference, producing notable open-set text-guided transfer results. Extensive experiments in both image and text settings demonstrate that our proposed method outperforms the state-of-the-art in both qualitative and quantitative evaluations.

Qinwei Lin, Xiaopeng Sun, Yu Gao et al.

Diffusion models have recently achieved outstanding results in the field of image super-resolution. These methods typically inject low-resolution (LR) images via ControlNet.In this paper, we first explore the temporal dynamics of information infusion through ControlNet, revealing that the input from LR images predominantly influences the initial stages of the denoising process. Leveraging this insight, we introduce a novel timestep-aware diffusion model that adaptively integrates features from both ControlNet and the pre-trained Stable Diffusion (SD). Our method enhances the transmission of LR information in the early stages of diffusion to guarantee image fidelity and stimulates the generation ability of the SD model itself more in the later stages to enhance the detail of generated images. To train this method, we propose a timestep-aware training strategy that adopts distinct losses at varying timesteps and acts on disparate modules. Experiments on benchmark datasets demonstrate the effectiveness of our method. Code: https://github.com/SleepyLin/TASR

Xiaopeng Sun, Qinwei Lin, Yu Gao et al.

Generative diffusion models (DM) have been extensively utilized in image super-resolution (ISR). Most of the existing methods adopt the denoising loss from DDPMs for model optimization. We posit that introducing reward feedback learning to finetune the existing models can further improve the quality of the generated images. In this paper, we propose a timestep-aware training strategy with reward feedback learning. Specifically, in the initial denoising stages of ISR diffusion, we apply low-frequency constraints to super-resolution (SR) images to maintain structural stability. In the later denoising stages, we use reward feedback learning to improve the perceptual and aesthetic quality of the SR images. In addition, we incorporate Gram-KL regularization to alleviate stylization caused by reward hacking. Our method can be integrated into any diffusion-based ISR model in a plug-and-play manner. Experiments show that ISR diffusion models, when fine-tuned with our method, significantly improve the perceptual and aesthetic quality of SR images, achieving excellent subjective results. Code: https://github.com/sxpro/RFSR

Yingsen Zeng, Zepeng Huang, Yujie Zhong et al.

Despite advances in general video understanding, Video Large Language Models (Video-LLMs) face challenges in precise temporal localization due to discrete time representations and limited temporally aware datasets. Existing methods for temporal expression either conflate time with text-based numerical values, add a series of dedicated temporal tokens, or regress time using specialized temporal grounding heads. To address these issues, we introduce DisTime, a lightweight framework designed to enhance temporal comprehension in Video-LLMs. DisTime employs a learnable token to create a continuous temporal embedding space and incorporates a Distribution-based Time Decoder that generates temporal probability distributions, effectively mitigating boundary ambiguities and maintaining temporal continuity. Additionally, the Distribution-based Time Encoder re-encodes timestamps to provide time markers for Video-LLMs. To overcome temporal granularity limitations in existing datasets, we propose an automated annotation paradigm that combines the captioning capabilities of Video-LLMs with the localization expertise of dedicated temporal models. This leads to the creation of InternVid-TG, a substantial dataset with 1.25M temporally grounded events across 179k videos, surpassing ActivityNet-Caption by 55 times. Extensive experiments demonstrate that DisTime achieves state-of-the-art performance across benchmarks in three time-sensitive tasks while maintaining competitive performance in Video QA tasks. Code and data are released at https://github.com/josephzpng/DisTime.

Wentao Xiang, Haoxian Tan, Cong Wei et al.

Perception is a fundamental task in the field of computer vision, encompassing a diverse set of subtasks that can be systematically categorized into four distinct groups based on two dimensions: prediction type and instruction type. Notably, existing researches often focus solely on a limited subset of these potential combinations, which constrains their applicability and versatility across various contexts. In response to this challenge, we present MVP-LM, a Multi-granular and Versatile Perception framework incorporating Visual Large Language Model. Our framework is designed to integrate both word-based and sentence-based perception tasks alongside box and mask predictions within a single architecture. MVP-LM features an innovative multi-granularity decoder in conjunction with a CoT-inspired dataset unification strategy, enabling seamless supervised fine-tuning across a wide spectrum of tasks, including but not limited to panoptic segmentation, detection, grounding, and referring expression segmentation. Furthermore, we introduce a query enhancement strategy aimed at harnessing the decoding and generative capabilities inherent in VLLMs. Extensive experiments conducted across a range of benchmarks in both word-based and sentence-based perception tasks substantiate the efficacy of our framework. The code will be available at https://github.com/xiangwentao666/MVP-LM.

Zelu Deng, Yujie Zhong, Sheng Guo et al.

This work aims at improving instance retrieval with self-supervision. We find that fine-tuning using the recently developed self-supervised (SSL) learning methods, such as SimCLR and MoCo, fails to improve the performance of instance retrieval. In this work, we identify that the learnt representations for instance retrieval should be invariant to large variations in viewpoint and background etc., whereas self-augmented positives applied by the current SSL methods can not provide strong enough signals for learning robust instance-level representations. To overcome this problem, we propose InsCLR, a new SSL method that builds on the \textit{instance-level} contrast, to learn the intra-class invariance by dynamically mining meaningful pseudo positive samples from both mini-batches and a memory bank during training. Extensive experiments demonstrate that InsCLR achieves similar or even better performance than the state-of-the-art SSL methods on instance retrieval. Code is available at https://github.com/zeludeng/insclr.

Chengjian Feng, Yujie Zhong, Yu Gao et al.

One-stage object detection is commonly implemented by optimizing two sub-tasks: object classification and localization, using heads with two parallel branches, which might lead to a certain level of spatial misalignment in predictions between the two tasks. In this work, we propose a Task-aligned One-stage Object Detection (TOOD) that explicitly aligns the two tasks in a learning-based manner. First, we design a novel Task-aligned Head (T-Head) which offers a better balance between learning task-interactive and task-specific features, as well as a greater flexibility to learn the alignment via a task-aligned predictor. Second, we propose Task Alignment Learning (TAL) to explicitly pull closer (or even unify) the optimal anchors for the two tasks during training via a designed sample assignment scheme and a task-aligned loss. Extensive experiments are conducted on MS-COCO, where TOOD achieves a 51.1 AP at single-model single-scale testing. This surpasses the recent one-stage detectors by a large margin, such as ATSS (47.7 AP), GFL (48.2 AP), and PAA (49.0 AP), with fewer parameters and FLOPs. Qualitative results also demonstrate the effectiveness of TOOD for better aligning the tasks of object classification and localization. Code is available at https://github.com/fcjian/TOOD.

Chengjian Feng, Yujie Zhong, Weilin Huang

The conventional detectors tend to make imbalanced classification and suffer performance drop, when the distribution of the training data is severely skewed. In this paper, we propose to use the mean classification score to indicate the classification accuracy for each category during training. Based on this indicator, we balance the classification via an Equilibrium Loss (EBL) and a Memory-augmented Feature Sampling (MFS) method. Specifically, EBL increases the intensity of the adjustment of the decision boundary for the weak classes by a designed score-guided loss margin between any two classes. On the other hand, MFS improves the frequency and accuracy of the adjustment of the decision boundary for the weak classes through over-sampling the instance features of those classes. Therefore, EBL and MFS work collaboratively for finding the classification equilibrium in long-tailed detection, and dramatically improve the performance of tail classes while maintaining or even improving the performance of head classes. We conduct experiments on LVIS using Mask R-CNN with various backbones including ResNet-50-FPN and ResNet-101-FPN to show the superiority of the proposed method. It improves the detection performance of tail classes by 15.6 AP, and outperforms the most recent long-tailed object detectors by more than 1 AP. Code is available at https://github.com/fcjian/LOCE.

Guanting Liu, Yujie Zhong, Sheng Guo et al.

Differentiable architecture search (DAS) has made great progress in searching for high-performance architectures with reduced computational cost. However, DAS-based methods mainly focus on searching for a repeatable cell structure, which is then stacked sequentially in multiple stages to form the networks. This configuration significantly reduces the search space, and ignores the importance of connections between the cells. To overcome this limitation, in this paper, we propose a Hierarchical Differentiable Architecture Search (H-DAS) that performs architecture search both at the cell level and at the stage level. Specifically, the cell-level search space is relaxed so that the networks can learn stage-specific cell structures. For the stage-level search, we systematically study the architectures of stages, including the number of cells in each stage and the connections between the cells. Based on insightful observations, we design several search rules and losses, and mange to search for better stage-level architectures. Such hierarchical search space greatly improves the performance of the networks without introducing expensive search cost. Extensive experiments on CIFAR10 and ImageNet demonstrate the effectiveness of the proposed H-DAS. Moreover, the searched stage-level architectures can be combined with the cell structures searched by existing DAS methods to further boost the performance. Code is available at: https://github.com/MalongTech/research-HDAS

Yujie Zhong, Linhai Xie, Sen Wang et al.

In this paper, we teach machines to understand visuals and natural language by learning the mapping between sentences and noisy video snippets without explicit annotations. Firstly, we define a self-supervised learning framework that captures the cross-modal information. A novel adversarial learning module is then introduced to explicitly handle the noises in the natural videos, where the subtitle sentences are not guaranteed to be strongly corresponded to the video snippets. For training and evaluation, we contribute a new dataset `ApartmenTour' that contains a large number of online videos and subtitles. We carry out experiments on the bidirectional retrieval tasks between sentences and videos, and the results demonstrate that our proposed model achieves the state-of-the-art performance on both retrieval tasks and exceeds several strong baselines. The dataset can be downloaded at https://github.com/zyj-13/WAL.

Chengjian Feng, Yujie Zhong, Zequn Jie et al.

In this paper, we present a novel paradigm to enhance the ability of object detector, e.g., expanding categories or improving detection performance, by training on synthetic dataset generated from diffusion models. Specifically, we integrate an instance-level grounding head into a pre-trained, generative diffusion model, to augment it with the ability of localising instances in the generated images. The grounding head is trained to align the text embedding of category names with the regional visual feature of the diffusion model, using supervision from an off-the-shelf object detector, and a novel self-training scheme on (novel) categories not covered by the detector. We conduct thorough experiments to show that, this enhanced version of diffusion model, termed as InstaGen, can serve as a data synthesizer, to enhance object detectors by training on its generated samples, demonstrating superior performance over existing state-of-the-art methods in open-vocabulary (+4.5 AP) and data-sparse (+1.2 to 5.2 AP) scenarios. Project page with code: https://fcjian.github.io/InstaGen.

Yu Gao, Jiancheng Huang, Xiaopeng Sun et al.

In this paper, we introduce Matten, a cutting-edge latent diffusion model with Mamba-Attention architecture for video generation. With minimal computational cost, Matten employs spatial-temporal attention for local video content modeling and bidirectional Mamba for global video content modeling. Our comprehensive experimental evaluation demonstrates that Matten has competitive performance with the current Transformer-based and GAN-based models in benchmark performance, achieving superior FVD scores and efficiency. Additionally, we observe a direct positive correlation between the complexity of our designed model and the improvement in video quality, indicating the excellent scalability of Matten.

Yan Fang, Mengcheng Lan, Zilong Huang et al.

In this paper, we present \textbf{Gen}erative \textbf{L}anguage-\textbf{I}mage \textbf{P}re-training (GenLIP), a minimalist generative pretraining framework for Vision Transformers (ViTs) designed for multimodal large language models (MLLMs). To better align vision encoders with the autoregressive nature of LLMs, GenLIP trains a ViT to predict language tokens directly from visual tokens using a standard language modeling objective, without contrastive batch construction or an additional text decoder. This design offers three key advantages: (1) \textbf{Simplicity}: a single transformer jointly models visual and textual tokens; (2) \textbf{Scalability}: it scales effectively with both data and model size; and (3) \textbf{Performance}: it achieves competitive or superior results across diverse multimodal benchmarks. Trained on 8B samples from Recap-DataComp-1B, GenLIP matches or surpasses strong baselines despite using substantially less pretraining data. After continued pretraining on multi-resolution images at native aspect ratios, GenLIP further improves on detail-sensitive tasks such as OCR and chart understanding, making it a strong foundation for vision encoders in MLLMs.

Lishuai Gao, Yujie Zhong, Yingsen Zeng et al.

Large Language Models (LLMs) have been widely used in various tasks, motivating us to develop an LLM-based assistant for videos. Instead of training from scratch, we propose a module to transform arbitrary well-trained image-based LLMs into video-LLMs (after being trained on video data). To better adapt image-LLMs for processing videos, we introduce two design principles: linear transformation to preserve the original visual-language alignment and representative information condensation from redundant video content. Guided by these principles, we propose a plug-and-play Linear Video Tokenizer(LinVT), which enables existing image-LLMs to understand videos. We benchmark LinVT with six recent visual LLMs: Aquila, Blip-3, InternVL2, Mipha, Molmo and Qwen2-VL, showcasing the high compatibility of LinVT. LinVT-based LLMs achieve state-of-the-art performance across various video benchmarks, illustrating the effectiveness of LinVT in multi-modal video understanding.

Lichen Ma, Tiezhu Yue, Pei Fu et al.

Recently, significant advancements have been made in diffusion-based visual text generation models. Although the effectiveness of these methods in visual text rendering is rapidly improving, they still encounter challenges such as inaccurate characters and strokes when rendering complex visual text. In this paper, we propose CharGen, a highly accurate character-level visual text generation and editing model. Specifically, CharGen employs a character-level multimodal encoder that not only extracts character-level text embeddings but also encodes glyph images character by character. This enables it to capture fine-grained cross-modality features more effectively. Additionally, we introduce a new perceptual loss in CharGen to enhance character shape supervision and address the issue of inaccurate strokes in generated text. It is worth mentioning that CharGen can be integrated into existing diffusion models to generate visual text with high accuracy. CharGen significantly improves text rendering accuracy, outperforming recent methods in public benchmarks such as AnyText-benchmark and MARIO-Eval, with improvements of more than 8% and 6%, respectively. Notably, CharGen achieved a 5.5% increase in accuracy on Chinese test sets.

Siyu Chen, Dengjie Li, Zenghao Bao et al.

Generating comics through text is widely studied. However, there are few studies on generating multi-panel Manga (Japanese comics) solely based on plain text. Japanese manga contains multiple panels on a single page, with characteristics such as coherence in storytelling, reasonable and diverse page layouts, consistency in characters, and semantic correspondence between panel drawings and panel scripts. Therefore, generating manga poses a significant challenge. This paper presents the manga generation task and constructs the Manga109Story dataset for studying manga generation solely from plain text. Additionally, we propose MangaDiffusion to facilitate the intra-panel and inter-panel information interaction during the manga generation process. The results show that our method particularly ensures the number of panels, reasonable and diverse page layouts. Based on our approach, there is potential to converting a large amount of textual stories into more engaging manga readings, leading to significant application prospects.

Baihui Xiao, Chengjian Feng, Zhijian Huang et al.

Collecting real-world data for rare high-risk scenarios, long-tailed driving events, and complex interactions remains challenging, leading to poor performance of existing autonomous driving systems in these critical situations. In this paper, we propose RoboTron-Sim that improves real-world driving in critical situations by utilizing simulated hard cases. First, we develop a simulated dataset called Hard-case Augmented Synthetic Scenarios (HASS), which covers 13 high-risk edge-case categories, as well as balanced environmental conditions such as day/night and sunny/rainy. Second, we introduce Scenario-aware Prompt Engineering (SPE) and an Image-to-Ego Encoder (I2E Encoder) to enable multimodal large language models to effectively learn real-world challenging driving skills from HASS, via adapting to environmental deviations and hardware differences between real-world and simulated scenarios. Extensive experiments on nuScenes show that RoboTron-Sim improves driving performance in challenging scenarios by around 50%, achieving state-of-the-art results in real-world open-loop planning. Qualitative results further demonstrate the effectiveness of RoboTron-Sim in better managing rare high-risk driving scenarios. Project page: https://stars79689.github.io/RoboTron-Sim/

Chi Zhang, Chengjian Feng, Feng Yan et al.

Video editing according to instructions is a highly challenging task due to the difficulty in collecting large-scale, high-quality edited video pair data. This scarcity not only limits the availability of training data but also hinders the systematic exploration of model architectures and training strategies. While prior work has improved specific aspects of video editing (e.g., synthesizing a video dataset using image editing techniques or decomposed video editing training), a holistic framework addressing the above challenges remains underexplored. In this study, we introduce InstructVEdit, a full-cycle instructional video editing approach that: (1) establishes a reliable dataset curation workflow to initialize training, (2) incorporates two model architectural improvements to enhance edit quality while preserving temporal consistency, and (3) proposes an iterative refinement strategy leveraging real-world data to enhance generalization and minimize train-test discrepancies. Extensive experiments show that InstructVEdit achieves state-of-the-art performance in instruction-based video editing, demonstrating robust adaptability to diverse real-world scenarios. Project page: https://o937-blip.github.io/InstructVEdit.

Dengjie Li, Tiancheng Shen, Yao Zhou et al.

Large language models (LLMs) have demonstrated remarkable capabilities, yet prohibitive parameter complexity often hinders their deployment. Existing singular value decomposition (SVD) based compression methods simply deem singular values as importance scores of decomposed components. However, this importance ordered by singular values does not necessarily correlate with the performance of a downstream task. In this work, we introduce SoCo (Singular spectrum optimization for large language model Compression), a novel compression framework that learns to rescale the decomposed components of SVD in a data-driven manner. Concretely, we employ a learnable diagonal matrix to assign importance scores for singular spectrum and develop a three-stage training process that progressively refines these scores from initial coarse compression to fine-grained sparsification-thereby striking an effective balance between aggressive model compression and performance preservation. Thanks to the learnable singular spectrum, SoCo adaptively prunes components according to the sparsified importance scores, rather than relying on the fixed order of singular values. More importantly, the remaining components with amplified importance scores can compensate for the loss of the pruned ones. Experimental evaluations across multiple LLMs and benchmarks demonstrate that SoCo surpasses the state-of-the-art methods in model compression.

Changcheng Xiao, Qiong Cao, Yujie Zhong et al.

Referring multi-object tracking (RMOT) is an emerging cross-modal task that aims to locate an arbitrary number of target objects and maintain their identities referred by a language expression in a video. This intricate task involves the reasoning of linguistic and visual modalities, along with the temporal association of target objects. However, the seminal work employs only loose feature fusion and overlooks the utilization of long-term information on tracked objects. In this study, we introduce a compact Transformer-based method, termed TenRMOT. We conduct feature fusion at both encoding and decoding stages to fully exploit the advantages of Transformer architecture. Specifically, we incrementally perform cross-modal fusion layer-by-layer during the encoding phase. In the decoding phase, we utilize language-guided queries to probe memory features for accurate prediction of the desired objects. Moreover, we introduce a query update module that explicitly leverages temporal prior information of the tracked objects to enhance the consistency of their trajectories. In addition, we introduce a novel task called Referring Multi-Object Tracking and Segmentation (RMOTS) and construct a new dataset named Ref-KITTI Segmentation. Our dataset consists of 18 videos with 818 expressions, and each expression averages 10.7 masks, which poses a greater challenge compared to the typical single mask in most existing referring video segmentation datasets. TenRMOT demonstrates superior performance on both the referring multi-object tracking and the segmentation tasks.

Chang-Bin Zhang, Jinhong Ni, Yujie Zhong et al.

In this paper, we address the challenging problem of open-world instance segmentation. Existing works have shown that vanilla visual networks are biased toward learning appearance information, \eg texture, to recognize objects. This implicit bias causes the model to fail in detecting novel objects with unseen textures in the open-world setting. To address this challenge, we propose a learning framework, called view-Consistent LeaRning (v-CLR), which aims to enforce the model to learn appearance-invariant representations for robust instance segmentation. In v-CLR, we first introduce additional views for each image, where the texture undergoes significant alterations while preserving the image's underlying structure. We then encourage the model to learn the appearance-invariant representation by enforcing the consistency between object features across different views, for which we obtain class-agnostic object proposals using off-the-shelf unsupervised models that possess strong object-awareness. These proposals enable cross-view object feature matching, greatly reducing the appearance dependency while enhancing the object-awareness. We thoroughly evaluate our method on public benchmarks under both cross-class and cross-dataset settings, achieving state-of-the-art performance. Project page: https://visual-ai.github.io/vclr

Xuange Zhang, Dengjie Li, Bo Liu et al.

Benefiting from recent advancements in large language models and modality alignment techniques, existing Large Vision-Language Models(LVLMs) have achieved prominent performance across a wide range of scenarios. However, the excessive computational complexity limits the widespread use of these models in practical applications. We argue that one main bottleneck in computational complexity is caused by the involvement of redundant vision sequences in model computation. This is inspired by a reassessment of the efficiency of vision and language information transmission in the language decoder of LVLMs. Then, we propose a novel hierarchical vision-language interaction mechanism called Hierarchical Vision injection for Mixture Attention (HiMix). In HiMix, only the language sequence undergoes full forward propagation, while the vision sequence interacts with the language at specific stages within each language decoder layer. It is striking that our approach significantly reduces computational complexity with minimal performance loss. Specifically, HiMix achieves a 10x reduction in the computational cost of the language decoder across multiple LVLM models while maintaining comparable performance. This highlights the advantages of our method, and we hope our research brings new perspectives to the field of vision-language understanding. Project Page: https://xuange923.github.io/HiMix

Chang-Bin Zhang, Yujie Zhong, Kai Han

Existing methods enhance the training of detection transformers by incorporating an auxiliary one-to-many assignment. In this work, we treat the model as a multi-task framework, simultaneously performing one-to-one and one-to-many predictions. We investigate the roles of each component in the transformer decoder across these two training targets, including self-attention, cross-attention, and feed-forward network. Our empirical results demonstrate that any independent component in the decoder can effectively learn both targets simultaneously, even when other components are shared. This finding leads us to propose a multi-route training mechanism, featuring a primary route for one-to-one prediction and two auxiliary training routes for one-to-many prediction. We propose a novel instructive self-attention mechanism, integrated into the first auxiliary route, which dynamically and flexibly guides object queries for one-to-many prediction. For the second auxiliary route, we introduce a route-aware Mixture-of-Experts (MoE) to facilitate knowledge sharing while mitigating potential conflicts between routes. Additionally, we apply an MoE to low-scale features in the encoder, optimizing the balance between efficiency and effectiveness. The auxiliary routes are discarded during inference. We conduct extensive experiments across various object detection baselines, achieving consistent improvements as demonstrated in Fig. 1. Our method is highly flexible and can be readily adapted to other tasks. To demonstrate its versatility, we conduct experiments on both instance segmentation and panoptic segmentation, further validating its effectiveness. Project page: https://visual-ai.github.io/mrdetr/

Lei Wang, Yujie Zhong, Xiaopeng Sun et al.

The task of 2D animal pose estimation plays a crucial role in advancing deep learning applications in animal behavior analysis and ecological research. Despite notable progress in some existing approaches, our study reveals that the scarcity of high-quality datasets remains a significant bottleneck, limiting the full potential of current methods. To address this challenge, we propose a novel Controllable Image Generation Pipeline for synthesizing animal pose estimation data, termed AP-CAP. Within this pipeline, we introduce a Multi-Modal Animal Image Generation Model capable of producing images with expected poses. To enhance the quality and diversity of the generated data, we further propose three innovative strategies: (1) Modality-Fusion-Based Animal Image Synthesis Strategy to integrate multi-source appearance representations, (2) Pose-Adjustment-Based Animal Image Synthesis Strategy to dynamically capture diverse pose variations, and (3) Caption-Enhancement-Based Animal Image Synthesis Strategy to enrich visual semantic understanding. Leveraging the proposed model and strategies, we create the MPCH Dataset (Modality-Pose-Caption Hybrid), the first hybrid dataset that innovatively combines synthetic and real data, establishing the largest-scale multi-source heterogeneous benchmark repository for animal pose estimation to date. Extensive experiments demonstrate the superiority of our method in improving both the performance and generalization capability of animal pose estimators.

Feng Yan, Weixin Luo, Yujie Zhong et al.

Existing end-to-end Multi-Object Tracking (e2e-MOT) methods have not surpassed non-end-to-end tracking-by-detection methods. One potential reason is its label assignment strategy during training that consistently binds the tracked objects with tracking queries and then assigns the few newborns to detection queries. With one-to-one bipartite matching, such an assignment will yield unbalanced training, i.e., scarce positive samples for detection queries, especially for an enclosed scene, as the majority of the newborns come on stage at the beginning of videos. Thus, e2e-MOT will be easier to yield a tracking terminal without renewal or re-initialization, compared to other tracking-by-detection methods. To alleviate this problem, we present Co-MOT, a simple and effective method to facilitate e2e-MOT by a novel coopetition label assignment with a shadow concept. Specifically, we add tracked objects to the matching targets for detection queries when performing the label assignment for training the intermediate decoders. For query initialization, we expand each query by a set of shadow counterparts with limited disturbance to itself. With extensive ablations, Co-MOT achieves superior performance without extra costs, e.g., 69.4% HOTA on DanceTrack and 52.8% TETA on BDD100K. Impressively, Co-MOT only requires 38\% FLOPs of MOTRv2 to attain a similar performance, resulting in the 1.4$\times$ faster inference speed.

Chengjian Feng, Yujie Zhong, Zequn Jie et al.

The goal of this work is to establish a scalable pipeline for expanding an object detector towards novel/unseen categories, using zero manual annotations. To achieve that, we make the following four contributions: (i) in pursuit of generalisation, we propose a two-stage open-vocabulary object detector, where the class-agnostic object proposals are classified with a text encoder from pre-trained visual-language model; (ii) To pair the visual latent space (of RPN box proposals) with that of the pre-trained text encoder, we propose the idea of regional prompt learning to align the textual embedding space with regional visual object features; (iii) To scale up the learning procedure towards detecting a wider spectrum of objects, we exploit the available online resource via a novel self-training framework, which allows to train the proposed detector on a large corpus of noisy uncurated web images. Lastly, (iv) to evaluate our proposed detector, termed as PromptDet, we conduct extensive experiments on the challenging LVIS and MS-COCO dataset. PromptDet shows superior performance over existing approaches with fewer additional training images and zero manual annotations whatsoever. Project page with code: https://fcjian.github.io/promptdet.