[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-layumi--Vehicle_reID-Collection":3,"tool-layumi--Vehicle_reID-Collection":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",149489,2,"2026-04-10T11:32:46",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",108322,"2026-04-10T11:39:34",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},6121,"gemini-cli","google-gemini\u002Fgemini-cli","gemini-cli 是一款由谷歌推出的开源 AI 命令行工具，它将强大的 Gemini 大模型能力直接集成到用户的终端环境中。对于习惯在命令行工作的开发者而言，它提供了一条从输入提示词到获取模型响应的最短路径，无需切换窗口即可享受智能辅助。\n\n这款工具主要解决了开发过程中频繁上下文切换的痛点，让用户能在熟悉的终端界面内直接完成代码理解、生成、调试以及自动化运维任务。无论是查询大型代码库、根据草图生成应用，还是执行复杂的 Git 操作，gemini-cli 都能通过自然语言指令高效处理。\n\n它特别适合广大软件工程师、DevOps 人员及技术研究人员使用。其核心亮点包括支持高达 100 万 token 的超长上下文窗口，具备出色的逻辑推理能力；内置 Google 搜索、文件操作及 Shell 命令执行等实用工具；更独特的是，它支持 MCP（模型上下文协议），允许用户灵活扩展自定义集成，连接如图像生成等外部能力。此外，个人谷歌账号即可享受免费的额度支持，且项目基于 Apache 2.0 协议完全开源，是提升终端工作效率的理想助手。",100752,"2026-04-10T01:20:03",[52,13,15,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":32,"last_commit_at":59,"category_tags":60,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":77,"owner_email":78,"owner_twitter":79,"owner_website":80,"owner_url":81,"languages":79,"stars":82,"forks":83,"last_commit_at":84,"license":79,"difficulty_score":85,"env_os":86,"env_gpu":87,"env_ram":87,"env_deps":88,"category_tags":91,"github_topics":92,"view_count":32,"oss_zip_url":79,"oss_zip_packed_at":79,"status":17,"created_at":104,"updated_at":105,"faqs":106,"releases":107},6211,"layumi\u002FVehicle_reID-Collection","Vehicle_reID-Collection",":red_car: the collection of vehicle re-ID papers, datasets. :red_car:","Vehicle_reID-Collection 是一个专注于车辆重识别（Vehicle Re-ID）领域的开源资源汇总库，旨在为相关研究提供一站式的论文、代码与数据集索引。在智能交通监控中，如何在不同摄像头视角下准确追踪同一辆车是一项极具挑战的任务，而该集合通过整理前沿算法与基准数据，有效降低了研究人员复现成果和开展新实验的门槛。\n\n这份资源特别适合计算机视觉领域的科研人员、算法工程师以及高校学生使用。其核心亮点在于“代码优先”的收录原则，重点聚合了带有公开实现代码的高质量论文，包括多次在 AICity Challenge 等国际竞赛中夺冠的解决方案。此外，它还涵盖了从经典的 VeRi-776、PKU Vehicle-ID 到大规模 VERI-Wild 等十余个主流数据集的详细链接与参数说明，并持续更新如基于 Transformer 的 TransReID 等最新技术架构。无论是希望快速搭建基线模型的开发者，还是致力于探索跨模态生成等前沿方向的学者，Vehicle_reID-Collection 都能提供坚实的数据支撑与技术参考，助力推动车辆识别技术的落地与应用。","# Vehicle Re-ID Collection\n\nIf you notice any result or the public code that has not been included in this page, please connect [Zhedong Zheng](mailto:zdzheng12@gmail.com) without hesitation to add the method. You are welcomed! \nor create pull request.\n\nPriorities are given to papers whose codes are published.\n\n## Code \n🏎️. The 1st Place Submission to AICity Challenge 2021 nlp re-id track (CVPR 2021 workshop) [[code]](https:\u002F\u002Fgithub.com\u002FShuaiBai623\u002FAIC2021-T5-CLV)[[paper]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FNLP-AICity2021\u002Fblob\u002Fmain\u002Fdoc\u002FCVPRW2021_NLP_AICity.pdf)\n\n🚙: The 2nd Place Submission to AICity Challenge 2021 re-id track (CVPR 2021 workshop) [[code]](https:\u002F\u002Fgithub.com\u002FXuanmeng-Zhang\u002FAICITY2021-Track2)\n\n:red_car:  The 1st Place Submission to AICity Challenge 2020 re-id track (CVPR 2020 workshop) [[code]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FAICIty-reID-2020)\n [[paper]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FAICIty-reID-2020\u002Fblob\u002Fmaster\u002Fpaper.pdf)\n \n :helicopter:  Drone-based building re-id (ACM Multimedia 2020) [[code]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FUniversity1652-Baseline)  [[paper]](https:\u002F\u002Farxiv.org\u002Fabs\u002F2002.12186)\n \n GPU-based Fast Re-Ranking [[code]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FPerson_reID_baseline_pytorch\u002Ftree\u002Fmaster\u002FGPU-Re-Ranking) [[paper]](https:\u002F\u002Farxiv.org\u002Fabs\u002F2012.07620v2)\n\n## Dataset\n1. VeRi-776\n\n    [project](https:\u002F\u002Fgithub.com\u002FVehicleReId\u002FVeRidataset) [paper](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-319-46475-6_53)\n\n49,357 images of 776 vehicles from 20 cameras. Like Market-1501 protocol.\n\nThe VeRi dataset is divided into a training subset containing 37,781 images of 576 subjects and a testing subset with 13,257 images of 200 subjects.then a query set containing 1,678 images of 200 subjects and a gallery including 11,579 image of 200 subjects are finally obtained.\n\n2. PKU Vehicle-ID\n\n    [project](https:\u002F\u002Fpkuml.org\u002Fresources\u002Fpku-vehicleid.html) [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016\u002Fpapers\u002FLiu_Deep_Relative_Distance_CVPR_2016_paper.pdf)\n\n221,763 images of 2,627 vehicles. Only two camera views??\n\n3. PKU-VD\n\n    [project](https:\u002F\u002Fpkuml.org\u002Fresources\u002Fpku-vds.html) [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FYan_Exploiting_Multi-Grain_Ranking_ICCV_2017_paper.pdf)\n\nwith attribute.\n\n4. VehicleReID\n\n    [project](https:\u002F\u002Fmedusa.fit.vutbr.cz\u002Ftraffic\u002Fresearch-topics\u002Fdetection-of-vehicles-and-datasets\u002Fvehicle-re-identification-for-automatic-video-traffic-surveillance-ats-cvpr-2016\u002F) [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016_workshops\u002Fw25\u002Fpapers\u002FZapletal_Vehicle_Re-Identification_for_CVPR_2016_paper.pdf)\n\n47,123 images from two cameras & lablled on pair.\n\n5. PKU-Vehicle\n\n    [project](http:\u002F\u002F59.110.216.11\u002Fhtml\u002F) [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8265213)\n\nno ID lablled.\n\n6. CompCars\n\n    [project](http:\u002F\u002Fmmlab.ie.cuhk.edu.hk\u002Fdatasets\u002Fcomp_cars\u002Findex.html) [pdf](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2015\u002Fpapers\u002FYang_A_Large-Scale_Car_2015_CVPR_paper.pdf) \n\n136,726 + 27,618 images of 1,716 cars with attributes. After crop,  136,713.\n\n7. StanfordCars\n\n    [project](http:\u002F\u002Fai.stanford.edu\u002F~jkrause\u002Fcars\u002Fcar_dataset.html) [pdf](http:\u002F\u002Fai.stanford.edu\u002F~jkrause\u002Fpapers\u002F3drr13.pdf)\n\n16,185 images of 196 classes.\n\n8. Vehicle-1M\n\n    [project](http:\u002F\u002Fwww.nlpr.ia.ac.cn\u002Fiva\u002Fhomepage\u002Fjqwang\u002FVehicle1M.htm) [pdf](https:\u002F\u002Fwww.aaai.org\u002Focs\u002Findex.php\u002FAAAI\u002FAAAI18\u002Fpaper\u002FviewFile\u002F16206\u002F16270)\n    \n    \n9. VERI-Wild \n\n[project](https:\u002F\u002Fgithub.com\u002FPKU-IMRE\u002FVERI-Wild)\n\n\n10. VRIC \nwith various motion blur and resolution\n\n[project](https:\u002F\u002Fqmul-vric.github.io) \n\n 60,430 images of 5,622 vehicle identities captured by 60 different cameras \n\n## Recent Papers\n\n### **2025**\n1. Coarse-to-Fine Cross-modality Generation for Enhancing Vehicle Re-Identification with High-Fidelity Synthetic Data **(ICRA)** [paper](https:\u002F\u002Fwww.zdzheng.xyz\u002Ffiles\u002FICRA25-Vehicle.pdf)\n\n### **2021**\n1. TransReID: Transformer-based Object Re-Identification **(ICCV)** \n[paper](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fpapers\u002FHe_TransReID_Transformer-Based_Object_Re-Identification_ICCV_2021_paper.pdf) \n[code](https:\u002F\u002Fgithub.com\u002Fdamo-cv\u002FTransReID)\n\n2. Viewpoint and Scale Consistency Reinforcement for UAV Vehicle Re-Identification **(IJCV)** \n[pdf](https:\u002F\u002Flink.springer.com\u002Fcontent\u002Fpdf\u002F10.1007\u002Fs11263-020-01402-2.pdf)\n\n3. Exploring Spatial Significance via Hybrid Pyramidal Graph Network for Vehicle Re-Identification **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9457192)\n\n4. PhD Learning: Learning With Pompeiu-Hausdorff Distances for Video-Based Vehicle Re-Identification **(CVPR)** \n[paper](http:\u002F\u002Fopenaccess.thecvf.com\u002F\u002Fcontent\u002FCVPR2021\u002Fhtml\u002FZhao_PhD_Learning_Learning_With_Pompeiu-Hausdorff_Distances_for_Video-Based_Vehicle_Re-Identification_CVPR_2021_paper.html) \n[code](https:\u002F\u002Fgithub.com\u002Femdata-ailab\u002FPhD-Learning)\n\n5. Heterogeneous Relational Complement for Vehicle Re-identification **(ICCV)** \n[paper](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fhtml\u002FZhao_Heterogeneous_Relational_Complement_for_Vehicle_Re-Identification_ICCV_2021_paper.html) \n[code](https:\u002F\u002Fgithub.com\u002FiCVTEAM\u002FHRCN)\n\n6. Model Latent Views With Multi-Center Metric Learning for Vehicle Re-Identification **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F9325909\u002F)\n\n7. Inter-Domain Adaptation Label for Data Augmentation in Vehicle Re-identification **(TMM)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9513554)\n\n8. Learning Multiple Semantic Knowledge For Cross-Domain Unsupervised Vehicle Re-Identification **(ICME)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9428440)\n\n9. Multi-level Deep Learning Vehicle Re-identification using Ranked-based Loss Functions **(ICPR)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9412415)\n\n10. Keypoint-Aligned Embeddings for Image Retrieval and Re-Identification **(WACV)** \n[pdf](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FWACV2021\u002Fpapers\u002FMoskvyak_Keypoint-Aligned_Embeddings_for_Image_Retrieval_and_Re-Identification_WACV_2021_paper.pdf)\n\n11. Pseudo Graph Convolutional Network for Vehicle ReID **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3474085.3475462)\n\n12. Vehicle Re-identification for Lane-level Travel Time Estimations on Congested Urban Road Networks Using Video Images **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9569748)\n\n13. OERFF: A Vehicle Re-Identification Method Based on Orientation Estimation and Regional Feature Fusion **(IEEE Access)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9416706)\n\n14. Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification **(ICCV)** \n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2108.08728) \n[code](https:\u002F\u002Fgithub.com\u002Fraoyongming\u002FCAL)\n\n15. Self-Supervised Geometric Features Discovery via Interpretable Attention for Vehicle Re-Identification and Beyond **(ICCV)** \n[pdf](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fpapers\u002FLi_Self-Supervised_Geometric_Features_Discovery_via_Interpretable_Attention_for_Vehicle_Re-Identification_ICCV_2021_paper.pdf)\n\n### **2020**\n1. VehicleNet: Learning Robust Visual Representation for Vehicle Re-identification **(TMM)**\n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.06305) \n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F186905783)\n\n2. Beyond the Parts: Learning Multi-view Cross-part Correlation for Vehicle Re-identification **(ACM MM)** \n[paper](http:\u002F\u002Fxinchenliu.com\u002Fpapers\u002F2020_ACMMM_PCRNet.pdf) \n[code](https:\u002F\u002Fgithub.com\u002Flxc86739795\u002Fparsing_platform)\n\n3. The Devil is in the Details: Self-Supervised Attention for Vehicle Re-Identification **(ECCV)** \n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.06271) \n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F191654655)\n\n4. Structural Analysis of Attributes for Vehicle Re-Identification and Retrieval **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8643580)\n\n5. Group-Group Loss-Based Global-Regional Feature Learning for Vehicle Re-Identification **(TIP)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8897720) \n[code](https:\u002F\u002Fgithub.com\u002Fliu-xb\u002FGGL)\n\n6. Simulating Content Consistent Vehicle Datasets with Attribute Descent **(ECCV)** \n[pdf](https:\u002F\u002Flink.springer.com\u002Fcontent\u002Fpdf\u002F10.1007\u002F978-3-030-58539-6_46.pdf) \n[code](https:\u002F\u002Fgithub.com\u002Fyorkeyao\u002FVehicleX)\n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F198061566)\n\n7. Parsing-based View-aware Embedding Network for Vehicle Re-Identification **(CVPR)** \n[paper](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_CVPR_2020\u002Fhtml\u002FMeng_Parsing-Based_View-Aware_Embedding_Network_for_Vehicle_Re-Identification_CVPR_2020_paper.html)\n[code](https:\u002F\u002Fgithub.com\u002Fsilverbulletmdc\u002FPVEN)\n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F160877803)\n\n8. Robust Re-Identification by Multiple Views Knowledge Distillation **(ECCV)** [paper](https:\u002F\u002Fwww.ecva.net\u002Fpapers\u002Feccv_2020\u002Fpapers_ECCV\u002Fhtml\u002F996_ECCV_2020_paper.php) \n[code](https:\u002F\u002Fgithub.com\u002Faimagelab\u002FVKD)\n\n9. Orientation-aware Vehicle Re-identification with Semantics-guided Part Attention Network **(ECCV)** \n[pdf](https:\u002F\u002Fpaperswithcode.com\u002Fpaper\u002Forientation-aware-vehicle-re-identification) \n[code](https:\u002F\u002Fgithub.com\u002Ftsaishien-chen\u002FSPAN)\n\n10. Vehicle Re-Identification Using Quadruple Directional Deep Learning Features **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8667847)\n\n11. Multi-Spectral Vehicle Re-Identification: A Challenge **(AAAI)** \n[paper](https:\u002F\u002Fojs.aaai.org\u002F\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F6796)\n\n12. Unsupervised Vehicle Re-identification with Progressive Adaptation **(IJCAI)** \n[paper](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2020\u002F127)\n\n13. Disentangled Feature Learning Network for Vehicle Re-Identification **(IJCAI)** \n[paper](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2020\u002F66)\n\n14. CFVMNet: A Multi-branch Network for Vehicle Re-identification Based on Common Field of View **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3394171.3413541)\n\n15. A Structured Graph Attention Network for Vehicle Re-Identification **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3394171.3413607)\n\n16. Fine-grained Feature Alignment with Part Perspective Transformation for Vehicle ReID **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3394171.3413573)\n\n17. Background Segmentation for Vehicle Re-identification **(MMM)** \n[paper](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-030-37734-2_8)\n\n18. Dual Domain Multi-Task Model for Vehicle Re-Identification **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9226133\u002F)\n\n19. Multi-View Spatial Attention Embedding for Vehicle Re-Identification **(TCSVT)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9033992)\n\n20. Unsupervised domain adaptive re-identification: Theory and practice **(PR)** \n[paper](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS003132031930473X)\n\n21. VARID: Viewpoint-Aware Re-IDentification of Vehicle Based on Triplet Loss **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9210535)\n\n22. Uncertainty-Aware Multi-Shot Knowledge Distillation for Image-Based Object Re-Identification **(AAAI)** \n[paper](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F6774)\n\n23. Tell The Truth From The Front: Anti-Disguise Vehicle Re-Identification **(ICME)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9102939)\n\n24. Vehicle Re-Identification Using Distance-Based Global and Partial Multi-Regional Feature Learning **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8972901)\n\n### **2019**\n1. VR-PROUD: Vehicle Re-identification using PROgressive Unsupervised Deep architecture **(PR)** [paper](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fabs\u002Fpii\u002FS0031320319300147)\n\n2. Embedding Adversarial Learning for Vehicle Re-Identification **(TIP)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8653852)\n\n3. Vehicle Re-Identification Using Quadruple Directional Deep Learning Features **(TITS)** [pdf](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.05163.pdf)\n\n4. VehicleNet: Learning Robust Feature Representation for Vehicle Re-identification **（CVPR workshop）** [paper](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_CVPRW_2019\u002Fhtml\u002FAI_City\u002FZheng_VehicleNet_Learning_Robust_Feature_Representation_for_Vehicle_Re-identification_CVPRW_2019_paper.html)\n\n5. Part-regularized Near-duplicate Vehicle Re-identification **(CVPR)** [pdf](http:\u002F\u002Fcvteam.net\u002Fpapers\u002F2019_CVPR_Part-regularized%20Near-duplicate%20Vehicle%20Re-identification.pdf)\n\n### **2018**\n1. Viewpoint-aware Attentive Multi-view Inference for Vehicle Re-identification **(CVPR)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018\u002Fpapers\u002FZhou_Viewpoint-Aware_Attentive_Multi-View_CVPR_2018_paper.pdf)\n\n2. Unsupervised Vehicle Re-Identification using Triplet Networks **(CVPR workshop)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018_workshops\u002Fpapers\u002Fw3\u002FMarin-Reyes_Unsupervised_Vehicle_Re-Identification_CVPR_2018_paper.pdf)\n\n3. Vehicle Re-Identification with the Space-Time Prior **(CVPR workshop)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018_workshops\u002Fpapers\u002Fw3\u002FWu_Vehicle_Re-Identification_With_CVPR_2018_paper.pdf)\n\n4. Fast vehicle identification via ranked semantic sampling based embedding **(IJCAI)** [pdf](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2018\u002F0514.pdf)\n\n5. Vehicle re-identification by deep hidden multi-view inference **(TIP)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8325486)\n\n6. Ram: a region-aware deep model for vehicle re-identification **(ICME)** [pdf](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.09283.pdf)\n\n7. Learning Coarse-to-Fine Structured Feature Embedding for Vehicle Re-Identification **(AAAI)** [pdf](https:\u002F\u002Fwww.aaai.org\u002Focs\u002Findex.php\u002FAAAI\u002FAAAI18\u002Fpaper\u002FviewFile\u002F16206\u002F16270)\n\n8. PROVID- Progressive and Multimodal Vehicle Reidentification for Large-Scale Urban Surveillance **(TMM)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8036238)\n\n9. Group Sensitive Triplet Embedding for Vehicle Re-identification **(TMM)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8265213)\n\n10. VP-ReID: vehicle and person re-identification system **(ACMMM)** [paper](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3206086)\n\n11. Vehicle Re-Identification by Adversarial Bi-Directional LSTM Network **(WACV)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8354181\u002F)\n\n12. Joint Semi-supervised Learning and Re-ranking for Vehicle Re-identification **(ICPR)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8545584\u002F)\n\n13. Multi-Attribute Driven Vehicle Re-Identification with Spatial-Temporal Re-Ranking **(ICIP)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8451776\u002F)\n\n14. Joint feature and similarity deep learning for vehicle re-identification **(IEEE Access)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8424333\u002F)\n### **2017**\n1. Orientation Invariant Feature Embedding and Spatial Temporal Regularization for Vehicle Re-Identification **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FWang_Orientation_Invariant_Feature_ICCV_2017_paper.pdf)\n\n2. Learning Deep Neural Networks for Vehicle Re-ID With Visual-Spatio-Temporal Path Proposals **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FShen_Learning_Deep_Neural_ICCV_2017_paper.pdf)\n\n3. Exploiting Multi-Grain Ranking Constraints for Precisely Searching Visually-similar Vehicles **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FYan_Exploiting_Multi-Grain_Ranking_ICCV_2017_paper.pdf)\n\n4. Improving triplet-wise training of convolutional neural network for vehicle re-identification **(ICME)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8019491)\n\n5. Deep hashing with multi-task learning for large-scale instance-level vehicle search **(ICME workshop)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8026274)\n\n6. Multi-modal metric learning for vehicle re-identification in traffic surveillance environment **(ICIP)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8296683)\n\n7. Vehicle re-identification by fusing multiple deep neural networks **(IPTA)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8310090)\n \n8. Beyond human-level license plate super-resolution with progressive vehicle search and domain priori GAN **(ACMMM)** [paper](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3123422)\n### **2016**\n1. Vehicle Re-Identification for Automatic Video Traffic Surveillance **(CVPR workshop)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016_workshops\u002Fw25\u002Fpapers\u002FZapletal_Vehicle_Re-Identification_for_CVPR_2016_paper.pdf)\n\n2. Deep Relative Distance Learning- Tell the Difference Between Similar Vehicles **(CVPR)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016\u002Fpapers\u002FLiu_Deep_Relative_Distance_CVPR_2016_paper.pdf)\n\n3. A Deep Learning-Based Approach to Progressive Vehicle Re-identification for Urban Surveillance **(ECCV)** [paper](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-319-46475-6_53)\n\n4. Large-Scale Vehicle Re-Identification in Urban Surveillance Videos **(ICME)** [pdf](https:\u002F\u002Fwww.researchgate.net\u002Fprofile\u002FXinchen_Liu\u002Fpublication\u002F303760492_Large-scale_vehicle_re-identification_in_urban_surveillance_videos\u002Flinks\u002F59e424090f7e9b97fbeb0ded\u002FLarge-scale-vehicle-re-identification-in-urban-surveillance-videos.pdf)\n\n### Reference\n- https:\u002F\u002Fgithub.com\u002Fbismex\u002FAwesome-vehicle-re-identification\n- https:\u002F\u002Fgithub.com\u002Fknwng\u002Fawesome-vehicle-re-identification\n","# 车辆重识别数据集与代码汇总\n\n如果您发现任何未在此页面列出的结果或公开代码，请随时联系 [Zhedong Zheng](mailto:zdzheng12@gmail.com)，以便将该方法添加进来。欢迎您的参与！或者直接创建拉取请求。\n\n优先考虑已公开代码的论文。\n\n## 代码 \n🏎️：2021年AICity挑战赛NLP重识别赛道冠军方案（CVPR 2021研讨会）[[代码]](https:\u002F\u002Fgithub.com\u002FShuaiBai623\u002FAIC2021-T5-CLV)[[论文]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FNLP-AICity2021\u002Fblob\u002Fmain\u002Fdoc\u002FCVPRW2021_NLP_AICity.pdf)\n\n🚙：2021年AICity挑战赛重识别赛道亚军方案（CVPR 2021研讨会）[[代码]](https:\u002F\u002Fgithub.com\u002FXuanmeng-Zhang\u002FAICITY2021-Track2)\n\n:red_car：2020年AICity挑战赛重识别赛道冠军方案（CVPR 2020研讨会）[[代码]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FAICIty-reID-2020)\n [[论文]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FAICIty-reID-2020\u002Fblob\u002Fmaster\u002Fpaper.pdf)\n \n :helicopter：基于无人机的建筑物重识别（ACM Multimedia 2020）[[代码]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FUniversity1652-Baseline)  [[论文]](https:\u002F\u002Farxiv.org\u002Fabs\u002F2002.12186)\n \n 基于GPU的快速重排序[[代码]](https:\u002F\u002Fgithub.com\u002Flayumi\u002FPerson_reID_baseline_pytorch\u002Ftree\u002Fmaster\u002FGPU-Re-Ranking) [[论文]](https:\u002F\u002Farxiv.org\u002Fabs\u002F2012.07620v2)\n\n## 数据集\n1. VeRi-776\n\n    [项目](https:\u002F\u002Fgithub.com\u002FVehicleReId\u002FVeRidataset) [论文](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-319-46475-6_53)\n\n包含来自20个摄像头的776辆车共49,357张图片。采用Market-1501协议。\n\nVeRi数据集分为训练子集和测试子集，其中训练子集包含576个目标的37,781张图片，测试子集包含200个目标的13,257张图片。随后进一步划分为查询集和画廊集，查询集包含200个目标的1,678张图片，画廊集则包含200个目标的11,579张图片。\n\n2. PKU Vehicle-ID\n\n    [项目](https:\u002F\u002Fpkuml.org\u002Fresources\u002Fpku-vehicleid.html) [PDF](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016\u002Fpapers\u002FLiu_Deep_Relative_Distance_CVPR_2016_paper.pdf)\n\n包含2,627辆车的221,763张图片。仅有两个视角的图像。\n\n3. PKU-VD\n\n    [项目](https:\u002F\u002Fpkuml.org\u002Fresources\u002Fpku-vds.html) [PDF](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FYan_Exploiting_Multi-Grain_Ranking_ICCV_2017_paper.pdf)\n\n带有属性信息。\n\n4. VehicleReID\n\n    [项目](https:\u002F\u002Fmedusa.fit.vutbr.cz\u002Ftraffic\u002Fresearch-topics\u002Fdetection-of-vehicles-and-datasets\u002Fvehicle-re-identification-for-automatic-video-traffic-surveillance-ats-cvpr-2016\u002F) [PDF](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016_workshops\u002Fw25\u002Fpapers\u002FZapletal_Vehicle_Re-Identification_for_CVPR_2016_paper.pdf)\n\n来自两台摄像机的47,123张图片，并进行了成对标注。\n\n5. PKU-Vehicle\n\n    [项目](http:\u002F\u002F59.110.216.11\u002Fhtml\u002F) [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8265213)\n\n未标注ID。\n\n6. CompCars\n\n    [项目](http:\u002F\u002Fmmlab.ie.cuhk.edu.hk\u002Fdatasets\u002Fcomp_cars\u002Findex.html) [PDF](https:\u002F\u002Fwww.cv-foundation.org\u002Fopenaccess\u002Fcontent_cvpr_2015\u002Fpapers\u002FYang_A_Large-Scale_Car_2015_CVPR_paper.pdf) \n\n包含1,716辆车的136,726 + 27,618张带属性的图片。裁剪后为136,713张。\n\n7. StanfordCars\n\n    [项目](http:\u002F\u002Fai.stanford.edu\u002F~jkrause\u002Fcars\u002Fcar_dataset.html) [PDF](http:\u002F\u002Fai.stanford.edu\u002F~jkrause\u002Fpapers\u002F3drr13.pdf)\n\n196个类别的16,185张图片。\n\n8. Vehicle-1M\n\n    [项目](http:\u002F\u002Fwww.nlpr.ia.ac.cn\u002Fiva\u002Fhomepage\u002Fjqwang\u002FVehicle1M.htm) [PDF](https:\u002F\u002Fwww.aaai.org\u002Focs\u002Findex.php\u002FAAAI\u002FAAAI18\u002Fpaper\u002FviewFile\u002F16206\u002F16270)\n    \n    \n9. VERI-Wild \n\n[项目](https:\u002F\u002Fgithub.com\u002FPKU-IMRE\u002FVERI-Wild)\n\n\n10. VRIC \n具有多种运动模糊和分辨率\n\n[项目](https:\u002F\u002Fqmul-vric.github.io) \n\n 60,430张图片，涵盖5,622个车辆身份，由60个不同摄像头拍摄\n\n## 最新论文\n\n### **2025**\n1. 基于高保真合成数据的粗粒度到细粒度跨模态生成以增强车辆重识别 **(ICRA)** [论文](https:\u002F\u002Fwww.zdzheng.xyz\u002Ffiles\u002FICRA25-Vehicle.pdf)\n\n### **2021年**\n1. TransReID：基于Transformer的目标重识别 **(ICCV)** \n[论文](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fpapers\u002FHe_TransReID_Transformer-Based_Object_Re-Identification_ICCV_2021_paper.pdf) \n[代码](https:\u002F\u002Fgithub.com\u002Fdamo-cv\u002FTransReID)\n\n2. 用于无人机车辆重识别的视角与尺度一致性增强 **(IJCV)** \n[PDF](https:\u002F\u002Flink.springer.com\u002Fcontent\u002Fpdf\u002F10.1007\u002Fs11263-020-01402-2.pdf)\n\n3. 基于混合金字塔图网络探索空间重要性以实现车辆重识别 **(TITS)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9457192)\n\n4. PhD学习：利用庞特尤-豪斯多夫距离进行基于视频的车辆重识别 **(CVPR)** \n[论文](http:\u002F\u002Fopenaccess.thecvf.com\u002F\u002Fcontent\u002FCVPR2021\u002Fhtml\u002FZhao_PhD_Learning_Learning_With_Pompeiu-Hausdorff_Distances_for_Video-Based_Vehicle_Re-Identification_CVPR_2021_paper.html) \n[代码](https:\u002F\u002Fgithub.com\u002Femdata-ailab\u002FPhD-Learning)\n\n5. 用于车辆重识别的异构关系互补 **(ICCV)** \n[论文](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fhtml\u002FZhao_Heterogeneous_Relational_Complement_for_Vehicle_Re-Identification_ICCV_2021_paper.html) \n[代码](https:\u002F\u002Fgithub.com\u002FiCVTEAM\u002FHRCN)\n\n6. 基于多中心度量学习对模型潜在视图进行建模以实现车辆重识别 **(TITS)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F9325909\u002F)\n\n7. 车辆重识别中用于数据增强的域间适应标签 **(TMM)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9513554)\n\n8. 学习多种语义知识以实现跨域无监督车辆重识别 **(ICME)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9428440)\n\n9. 基于排序损失函数的多级深度学习车辆重识别 **(ICPR)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9412415)\n\n10. 用于图像检索和重识别的关键点对齐嵌入 **(WACV)** \n[PDF](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FWACV2021\u002Fpapers\u002FMoskvyak_Keypoint-Aligned_Embeddings_for_Image_Retrieval_and_Re-Identification_WACV_2021_paper.pdf)\n\n11. 用于车辆重识别的伪图卷积网络 **(ACMMM)** \n[论文](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3474085.3475462)\n\n12. 基于视频图像的拥堵城市路网车道级行程时间估计中的车辆重识别 **(TITS)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9569748)\n\n13. OERFF：一种基于方向估计与区域特征融合的车辆重识别方法 **(IEEE Access)** \n[论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9416706)\n\n14. 用于细粒度视觉分类与重识别的反事实注意力学习 **(ICCV)** \n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2108.08728) \n[代码](https:\u002F\u002Fgithub.com\u002Fraoyongming\u002FCAL)\n\n15. 基于可解释注意力的自监督几何特征发现，用于车辆重识别及其他任务 **(ICCV)** \n[PDF](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent\u002FICCV2021\u002Fpapers\u002FLi_Self-Supervised_Geometric_Features_Discovery_via_Interpretable_Attention_for_Vehicle_Re-Identification_ICCV_2021_paper.pdf)\n\n### **2020年**\n1. VehicleNet：学习鲁棒的车辆重识别视觉表征 **(TMM)**\n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.06305) \n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F186905783)\n\n2. 不止于局部：学习多视角跨部件相关性用于车辆重识别 **(ACM MM)** \n[paper](http:\u002F\u002Fxinchenliu.com\u002Fpapers\u002F2020_ACMMM_PCRNet.pdf) \n[code](https:\u002F\u002Fgithub.com\u002Flxc86739795\u002Fparsing_platform)\n\n3. 细节决定成败：自监督注意力机制用于车辆重识别 **(ECCV)** \n[arXiv](https:\u002F\u002Farxiv.org\u002Fabs\u002F2004.06271) \n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F191654655)\n\n4. 基于属性的结构化分析用于车辆重识别与检索 **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8643580)\n\n5. 基于组-组损失的全局-区域特征学习用于车辆重识别 **(TIP)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8897720) \n[code](https:\u002F\u002Fgithub.com\u002Fliu-xb\u002FGGL)\n\n6. 利用属性下降法模拟内容一致的车辆数据集 **(ECCV)** \n[pdf](https:\u002F\u002Flink.springer.com\u002Fcontent\u002Fpdf\u002F10.1007\u002F978-3-030-58539-6_46.pdf) \n[code](https:\u002F\u002Fgithub.com\u002Fyorkeyao\u002FVehicleX)\n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F198061566)\n\n7. 基于解析的视图感知嵌入网络用于车辆重识别 **(CVPR)** \n[paper](https:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_CVPR_2020\u002Fhtml\u002FMeng_Parsing-Based_View-Aware_Embedding_Network_for_Vehicle_Re-Identification_CVPR_2020_paper.html)\n[code](https:\u002F\u002Fgithub.com\u002Fsilverbulletmdc\u002FPVEN)\n[[中文介绍]](https:\u002F\u002Fzhuanlan.zhihu.com\u002Fp\u002F160877803)\n\n8. 多视角知识蒸馏实现鲁棒的重识别 **(ECCV)** [paper](https:\u002F\u002Fwww.ecva.net\u002Fpapers\u002Feccv_2020\u002Fpapers_ECCV\u002Fhtml\u002F996_ECCV_2020_paper.php) \n[code](https:\u002F\u002Fgithub.com\u002Faimagelab\u002FVKD)\n\n9. 基于语义引导的部件注意力网络的朝向感知车辆重识别 **(ECCV)** \n[pdf](https:\u002F\u002Fpaperswithcode.com\u002Fpaper\u002Forientation-aware-vehicle-re-identification) \n[code](https:\u002F\u002Fgithub.com\u002Ftsaishien-chen\u002FSPAN)\n\n10. 利用四重方向深度学习特征进行车辆重识别 **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fdocument\u002F8667847)\n\n11. 多光谱车辆重识别：一项挑战 **(AAAI)** \n[paper](https:\u002F\u002Fojs.aaai.org\u002F\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F6796)\n\n12. 无监督车辆重识别与渐进式适应 **(IJCAI)** \n[paper](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2020\u002F127)\n\n13. 用于车辆重识别的解耦特征学习网络 **(IJCAI)** \n[paper](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2020\u002F66)\n\n14. CFVMNet：基于共同视场的多分支车辆重识别网络 **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3394171.3413541)\n\n15. 用于车辆重识别的结构化图注意力网络 **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002F10.1145\u002F3394171.3413607)\n\n16. 基于部件视角变换的细粒度特征对齐用于车辆重识别 **(ACMMM)** \n[paper](https:\u002F\u002Fdl.acm.org\u002Fdoi\u002Fabs\u002F10.1145\u002F3394171.3413573)\n\n17. 背景分割用于车辆重识别 **(MMM)** \n[paper](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-030-37734-2_8)\n\n18. 双域多任务模型用于车辆重识别 **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9226133\u002F)\n\n19. 多视角空间注意力嵌入用于车辆重识别 **(TCSVT)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9033992)\n\n20. 无监督领域自适应重识别：理论与实践 **(PR)** \n[paper](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fpii\u002FS003132031930473X)\n\n21. VARID：基于三元组损失的视角感知车辆重识别 **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9210535)\n\n22. 不确定性感知的多样本知识蒸馏用于基于图像的目标重识别 **(AAAI)** \n[paper](https:\u002F\u002Fojs.aaai.org\u002Findex.php\u002FAAAI\u002Farticle\u002Fview\u002F6774)\n\n23. 从正面看清真相：防伪装车辆重识别 **(ICME)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F9102939)\n\n24. 利用基于距离的全局与局部多区域特征学习进行车辆重识别 **(TITS)** \n[paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8972901)\n\n### **2019年**\n1. VR-PROUD：使用渐进式无监督深度架构进行车辆重识别 **(PR)** [paper](https:\u002F\u002Fwww.sciencedirect.com\u002Fscience\u002Farticle\u002Fabs\u002Fpii\u002FS0031320319300147)\n\n2. 嵌入对抗学习用于车辆重识别 **(TIP)** [paper](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8653852)\n\n3. 利用四重方向深度学习特征进行车辆重识别 **(TITS)** [pdf](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1811.05163.pdf)\n\n4. VehicleNet：学习鲁棒的车辆重识别特征表示 **（CVPR研讨会）** [paper](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_CVPRW_2019\u002Fhtml\u002FAI_City\u002FZheng_VehicleNet_Learning_Robust_Feature_Representation_for_Vehicle_Re-identification_CVPRW_2019_paper.html)\n\n5. 部件正则化的近似重复车辆重识别 **(CVPR)** [pdf](http:\u002F\u002Fcvteam.net\u002Fpapers\u002F2019_CVPR_Part-regularized%20Near-duplicate%20Vehicle%20Re-identification.pdf)\n\n### **2018年**\n1. 基于视点感知的多视角注意力推理用于车辆重识别 **(CVPR)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018\u002Fpapers\u002FZhou_Viewpoint-Aware_Attentive_Multi-View_CVPR_2018_paper.pdf)\n\n2. 使用三元组网络的无监督车辆重识别 **(CVPR研讨会)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018_workshops\u002Fpapers\u002Fw3\u002FMarin-Reyes_Unsupervised_Vehicle_Re-Identification_CVPR_2018_paper.pdf)\n\n3. 基于时空先验的车辆重识别 **(CVPR研讨会)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2018_workshops\u002Fpapers\u002Fw3\u002FWu_Vehicle_Re-Identification_With_CVPR_2018_paper.pdf)\n\n4. 基于排序语义采样嵌入的快速车辆识别 **(IJCAI)** [pdf](https:\u002F\u002Fwww.ijcai.org\u002Fproceedings\u002F2018\u002F0514.pdf)\n\n5. 通过深度隐式多视角推理进行车辆重识别 **(TIP)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8325486)\n\n6. Ram：一种区域感知的深度模型用于车辆重识别 **(ICME)** [pdf](https:\u002F\u002Farxiv.org\u002Fpdf\u002F1806.09283.pdf)\n\n7. 学习粗到细的结构化特征嵌入用于车辆重识别 **(AAAI)** [pdf](https:\u002F\u002Fwww.aaai.org\u002Focs\u002Findex.php\u002FAAAI\u002FAAAI18\u002Fpaper\u002FviewFile\u002F16206\u002F16270)\n\n8. PROVID：面向大规模城市监控的渐进式多模态车辆重识别 **(TMM)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8036238)\n\n9. 面向车辆重识别的群体敏感三元组嵌入 **(TMM)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8265213)\n\n10. VP-ReID：车辆与人员重识别系统 **(ACMMM)** [论文](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3206086)\n\n11. 基于对抗性双向LSTM网络的车辆重识别 **(WACV)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8354181\u002F)\n\n12. 车辆重识别中的联合半监督学习与重排序 **(ICPR)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8545584\u002F)\n\n13. 基于多属性驱动与时空重排序的车辆重识别 **(ICIP)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8451776\u002F)\n\n14. 车辆重识别中的联合特征与相似度深度学习 **(IEEE Access)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8424333\u002F)\n### **2017年**\n1. 用于车辆重识别的方向不变特征嵌入及时空正则化 **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FWang_Orientation_Invariant_Feature_ICCV_2017_paper.pdf)\n\n2. 基于视觉-时空路径提案的深度神经网络用于车辆重识别 **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FShen_Learning_Deep_Neural_ICCV_2017_paper.pdf)\n\n3. 利用多粒度排序约束精确搜索视觉相似车辆 **(ICCV)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_ICCV_2017\u002Fpapers\u002FYan_Exploiting_Multi-Grain_Ranking_ICCV_2017_paper.pdf)\n\n4. 改进卷积神经网络的三元组训练以用于车辆重识别 **(ICME)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8019491)\n\n5. 基于多任务学习的深度哈希用于大规模实例级车辆搜索 **(ICME研讨会)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8026274)\n\n6. 交通监控环境下用于车辆重识别的多模态度量学习 **(ICIP)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8296683)\n\n7. 通过融合多个深度神经网络进行车辆重识别 **(IPTA)** [论文](https:\u002F\u002Fieeexplore.ieee.org\u002Fabstract\u002Fdocument\u002F8310090)\n\n8. 超越人类水平的车牌超分辨率，结合渐进式车辆搜索与领域先验GAN **(ACMMM)** [论文](https:\u002F\u002Fdl.acm.org\u002Fcitation.cfm?id=3123422)\n### **2016年**\n1. 用于自动视频交通监控的车辆重识别 **(CVPR研讨会)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016_workshops\u002Fw25\u002Fpapers\u002FZapletal_Vehicle_Re-Identification_for_CVPR_2016_paper.pdf)\n\n2. 深度相对距离学习——区分相似车辆 **(CVPR)** [pdf](http:\u002F\u002Fopenaccess.thecvf.com\u002Fcontent_cvpr_2016\u002Fpapers\u002FLiu_Deep_Relative_Distance_CVPR_2016_paper.pdf)\n\n3. 基于深度学习的渐进式车辆重识别方法用于城市监控 **(ECCV)** [论文](https:\u002F\u002Flink.springer.com\u002Fchapter\u002F10.1007\u002F978-3-319-46475-6_53)\n\n4. 城市监控视频中的大规模车辆重识别 **(ICME)** [pdf](https:\u002F\u002Fwww.researchgate.net\u002Fprofile\u002FXinchen_Liu\u002Fpublication\u002F303760492_Large-scale_vehicle_re-identification_in_urban_surveillance_videos\u002Flinks\u002F59e424090f7e9b97fbeb0ded\u002FLarge-scale-vehicle-re-identification-in-urban-surveillance-videos.pdf)\n\n### 参考文献\n- https:\u002F\u002Fgithub.com\u002Fbismex\u002FAwesome-vehicle-re-identification\n- https:\u002F\u002Fgithub.com\u002Fknwng\u002Fawesome-vehicle-re-identification","# Vehicle_reID-Collection 快速上手指南\n\nVehicle_reID-Collection 是一个汇集车辆重识别（Vehicle Re-ID）领域顶尖代码、数据集和最新论文的开源资源库。本指南将帮助你快速了解如何利用该集合中的资源开始开发。\n\n## 环境准备\n\n在开始之前，请确保你的开发环境满足以下基本要求。由于该集合包含多个独立项目，具体依赖可能略有不同，但通用环境如下：\n\n*   **操作系统**: Linux (推荐 Ubuntu 18.04\u002F20.04) 或 macOS。Windows 用户建议使用 WSL2。\n*   **Python**: 3.6 或更高版本 (推荐 3.8+)。\n*   **深度学习框架**: PyTorch (大多数项目基于此，版本需与具体子项目匹配，通常建议 1.7+)。\n*   **GPU**: 推荐使用 NVIDIA GPU 并安装对应的 CUDA 驱动及 Toolkit，以加速训练和推理（特别是涉及 Re-Ranking 和 Transformer 模型时）。\n*   **基础依赖库**:\n    ```bash\n    pip install numpy scipy matplotlib opencv-python tqdm scikit-image scikit-learn yacs\n    ```\n\n> **提示**：国内开发者建议使用清华源或阿里源加速安装：\n> `pip install -r requirements.txt -i https:\u002F\u002Fpypi.tuna.tsinghua.edu.cn\u002Fsimple`\n\n## 安装步骤\n\n由于 `Vehicle_reID-Collection` 本身是一个资源索引列表，而非单一的可安装包，你需要根据需求克隆具体的子项目代码。以下是获取热门冠军方案（如 AICity Challenge 2021 第一名）的通用步骤：\n\n1.  **克隆仓库**\n    选择你感兴趣的具体项目链接进行克隆。以 2021 年 AICity Challenge NLP Re-ID 赛道冠军代码为例：\n    ```bash\n    git clone https:\u002F\u002Fgithub.com\u002FShuaiBai623\u002FAIC2021-T5-CLV.git\n    cd AIC2021-T5-CLV\n    ```\n\n2.  **安装项目特定依赖**\n    进入目录后，检查并安装该项目独有的 `requirements.txt`：\n    ```bash\n    pip install -r requirements.txt\n    ```\n    *注：若项目中包含自定义 CUDA 算子（如 GPU-Re-Ranking），可能需要额外编译步骤，请参考该项目内的 README。*\n\n3.  **准备数据集**\n    下载所需数据集（如 VeRi-776, VehicleID 等）。\n    *   **VeRi-776**: 从 [项目主页](https:\u002F\u002Fgithub.com\u002FVehicleReId\u002FVeRidataset) 下载。\n    *   **目录结构**: 通常需要将数据解压并整理为以下格式（具体视项目要求而定）：\n        ```text\n        data\u002F\n        └── veri\u002F\n            ├── image_train\u002F\n            ├── image_test\u002F\n            └── query_test\u002F\n        ```\n\n## 基本使用\n\n以下以典型的 PyTorch 车辆重识别基线模型为例，展示训练和测试的最简流程。具体命令参数请以各子项目的官方文档为准。\n\n### 1. 训练模型 (Training)\n\n使用单张 GPU 启动训练，指定数据集路径和保存目录：\n\n```bash\npython train.py --data_dir .\u002Fdata\u002Fveri --save_dir .\u002Flogs\u002Fveri_baseline --gpu_ids 0\n```\n\n*   `--data_dir`: 指向预处理好的数据集根目录。\n*   `--save_dir`: 模型权重和日志的输出路径。\n*   `--gpu_ids`: 指定使用的 GPU 编号。\n\n### 2. 测试与评估 (Testing)\n\n训练完成后，使用生成的模型权重在查询集（Query）和图库集（Gallery）上进行特征提取和排名评估：\n\n```bash\npython test.py --data_dir .\u002Fdata\u002Fveri --checkpoint .\u002Flogs\u002Fveri_baseline\u002Fmodel_best.pth --re_ranking\n```\n\n*   `--checkpoint`: 指向训练好的 `.pth` 模型文件。\n*   `--re_ranking`: 启用重排序策略（如 k-reciprocal encoding），通常能显著提升 mAP 和 Rank-1 指标。该集合中包含了高效的 [GPU-based Fast Re-Ranking](https:\u002F\u002Fgithub.com\u002Flayumi\u002FPerson_reID_baseline_pytorch\u002Ftree\u002Fmaster\u002FGPU-Re-Ranking) 实现。\n\n### 3. 结果解读\n\n运行结束后，终端将输出关键评估指标：\n*   **mAP**: 平均精度均值。\n*   **Rank-1 \u002F Rank-5 \u002F Rank-10**: 正确匹配出现在前 N 位的概率。\n\n你可以参考集合中列出的 [Recent Papers](#recent-papers) 部分，对比你的结果与 SOTA（State-of-the-Art）方法的差距，并尝试复现如 **TransReID** 或 **VehicleNet** 等先进架构。","某智慧交通团队正在开发城市级车辆追踪系统，需要从海量监控视频中快速锁定并还原特定车辆的行驶轨迹。\n\n### 没有 Vehicle_reID-Collection 时\n- **数据搜集如大海捞针**：团队成员需花费数周时间在各大学术网站零星搜索，难以一次性获取 VeRi-776、VERI-Wild 等涵盖多摄像头视角的高质量数据集，导致模型训练数据匮乏且格式不统一。\n- **复现顶尖算法门槛高**：面对 AICity Challenge 等竞赛的冠军方案（如 2021 年第一名提交代码），因缺乏官方整理的开源实现链接，开发人员需从头构建基线，极易在环境配置和细节处理上踩坑。\n- **技术选型盲目低效**：由于缺少按年份和性能排序的论文清单（如从传统 CNN 到最新的 TransReID Transformer 架构），团队难以评估哪种算法最适合当前模糊或低分辨率的监控场景，试错成本极高。\n\n### 使用 Vehicle_reID-Collection 后\n- **一站式获取权威数据**：直接下载整理好的 VeRi-776 和 VRIC 等数据集，这些数据集已预处理好训练集与测试集划分，包含丰富的车辆属性标注，让模型训练立即启动。\n- **站在巨人肩膀上创新**：直接调用 GitHub 上经过验证的 AICity 竞赛冠军代码库（如 GPU 加速重排序模块），将原本需要一个月的基线搭建工作缩短至两天，并能快速集成最新的高保真合成数据生成技术。\n- **精准匹配前沿方案**：通过清晰的论文列表，迅速定位到针对运动模糊和低分辨率优化的最新研究（如 2025 年 ICRA 论文），针对性地解决了夜间或恶劣天气下的车辆识别难题。\n\nVehicle_reID-Collection 将原本分散杂乱的学术资源转化为标准化的工程资产，让研发团队能从繁琐的资源整理中解脱，专注于核心业务逻辑的突破。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Flayumi_Vehicle_reID-Collection_2c4118a0.png","layumi","Zhedong Zheng","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Flayumi_7083c170.jpg","Hi, I am a tenure-track assistant professor at the University of Macau. My work focuses on computer vision, especially representation learning. ","University of Macau","Macau, China","Zhedong.Zheng@student.uts.edu.au",null,"http:\u002F\u002Fwww.zdzheng.xyz","https:\u002F\u002Fgithub.com\u002Flayumi",510,55,"2026-03-24T12:00:51",4,"","未说明",{"notes":89,"python":87,"dependencies":90},"该 README 仅为车辆重识别（Vehicle Re-ID）相关论文、代码库链接及数据集的汇总列表，并非单一可执行工具的说明文档。文中列出的各个子项目（如 TransReID, VehicleNet 等）拥有各自独立的运行环境要求，需分别访问其对应的 GitHub 仓库查看具体的操作系统、GPU、Python 版本及依赖库信息。",[],[15,14,16],[93,94,95,96,97,98,99,100,101,102,103],"vehicle-reid","pku-vehicle","deep-learning","paper","ve-ri","veri776","vehicle","cvpr-workshop","dataset","awesome","awesome-list","2026-03-27T02:49:30.150509","2026-04-10T20:44:06.364902",[],[]]