行人重识别(Person re-identification)也称行人再识别,是利用计算机视觉技术判断图像或者视频序列中是否存在特定行人的技术。广泛被认为是一个图像检索的子问题。给定一个监控行人图像,检索跨设备下的该行人图像。旨在弥补目前固定的摄像头的视觉局限,并可与行人检测/行人跟踪技术相结合,可广泛应用于智能视频监控、智能安保等领域。 由于不同摄像设备之间的差异,同时行人兼具刚性和柔性的特性 ,外观易受穿着、尺

知识荟萃

行人重识别 Person Re-identification / Person Retrieval 专知荟萃

入门学习

  1. 行人重识别综述
  2. 基于深度学习的Person Re-ID(综述)
  3. 郑哲东 -Deep-ReID:行人重识别的深度学习方法
  4. 【行人识别】Deep Transfer Learning for Person Re-identification
  5. 知乎专栏:行人重识别 [https://zhuanlan.zhihu.com/personReid]
    • 行人重识别综述:从哈利波特地图说起
    • 行人再识别中的迁移学习:图像风格转换(Learning via Translation)
    • 行人对齐+重识别网络
    • SVDNet for Pedestrian Retrieval:CNN到底认为哪个投影方向是重要的?
    • 用GAN生成的图像做训练?Yes!
    • 2017 ICCV 行人检索/重识别 接受论文汇总
    • 从人脸识别 到 行人重识别,下一个风口
  6. GAN(生成式对抗网络)的研究现状,以及在行人重识别领域的应用前景?
  7. Re-id Resources
  8. 行人再识别(行人重识别)【包含与行人检测的对比】
  9.  行人重识别综述(Person Re-identification: Past, Present and Future)

综述

  1. Person Re-identification: Past, Present and Future Liang Zheng, Yi Yang, Alexander G. Hauptmann

  2. Person Re-Identification Book

  3. A Systematic Evaluation and Benchmark for Person Re-Identification: Features, Metrics, and Datasets

  4. People reidentification in surveillance and forensics: A survey

  5. 比较全的Paper List 集合:

进阶论文及代码

Person Re-identification / Person Retrieval

参考链接: [https://github.com/handong1587/handong1587.github.io/blob/master/_posts/deep_learning/2015-10-09-re-id.md]
  1. DeepReID: Deep Filter Pairing Neural Network for Person Re-Identification
  2. An Improved Deep Learning Architecture for Person Re-Identification
  3. Deep Ranking for Person Re-identification via Joint Representation Learning
  4. PersonNet: Person Re-identification with Deep Convolutional Neural Networks
  5. Learning Deep Feature Representations with Domain Guided Dropout for Person Re-identification
  6. Person Re-Identification by Multi-Channel Parts-Based CNN with Improved Triplet Loss Function
  7. End-to-End Comparative Attention Networks for Person Re-identification
  8. A Multi-task Deep Network for Person Re-identification
  9. Gated Siamese Convolutional Neural Network Architecture for Human Re-Identification
  10. A Siamese Long Short-Term Memory Architecture for Human Re-Identification
  11. Gated Siamese Convolutional Neural Network Architecture for Human Re-Identification
  12. Person Re-identification: Past, Present and Future
  13. Deep Learning Prototype Domains for Person Re-Identification
  14. Deep Transfer Learning for Person Re-identification
  15. A Discriminatively Learned CNN Embedding for Person Re-identification
  16. Structured Deep Hashing with Convolutional Neural Networks for Fast Person Re-identification
  17. In Defense of the Triplet Loss for Person Re-Identification
  18. Beyond triplet loss: a deep quadruplet network for person re-identification
  19. Part-based Deep Hashing for Large-scale Person Re-identification
  20. Deep Person Re-Identification with Improved Embedding
  21. Towards a Principled Integration of Multi-Camera Re-Identification and Tracking through Optimal Bayes Filters
  22. Person Re-Identification by Deep Joint Learning of Multi-Loss Classification
  23. Attention-based Natural Language Person Retrieval
  24. Unsupervised Person Re-identification: Clustering and Fine-tuning
  25. Deep Representation Learning with Part Loss for Person Re-Identification
  26. Pedestrian Alignment Network for Large-scale Person Re-identification
  27. Deep Reinforcement Learning Attention Selection for Person Re-Identification
  28. Learning Efficient Image Representation for Person Re-Identification
  29. Person Re-identification Using Visual Attention
  30. Deeply-Learned Part-Aligned Representations for Person Re-Identification
  31. What-and-Where to Match: Deep Spatially Multiplicative Integration Networks for Person Re-identification
  32. Deep Feature Learning via Structured Graph Laplacian Embedding for Person Re-Identification
  33. Divide and Fuse: A Re-ranking Approach for Person Re-identification
  34. Large Margin Learning in Set to Set Similarity Comparison for Person Re-identification
  35. Multi-scale Deep Learning Architectures for Person Re-identification
  36. Pose-driven Deep Convolutional Model for Person Re-identification
  37. HydraPlus-Net: Attentive Deep Features for Pedestrian Analysis
  38. Person Re-Identification with Vision and Language
  39. Margin Sample Mining Loss: A Deep Learning Based Method for Person Re-identification
  40. Learning Deep Context-aware Features over Body and Latent Parts for Person Re-identification
  41. Pseudo-positive regularization for deep person re-identification
  42. Let Features Decide for Themselves: Feature Mask Network for Person Re-identification
  43. Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification
  44. AlignedReID: Surpassing Human-Level Performance in Person Re-Identification
  45. Region-based Quality Estimation Network for Large-scale Person Re-identification
  46. Deep-Person: Learning Discriminative Deep Features for Person Re-Identification
  47. A Pose-Sensitive Embedding for Person Re-Identification with Expanded Cross Neighborhood Re-Ranking
  1. Joint Detection and Identification Feature Learning for Person Search

  2. Person Re-identification in the Wild

  3. IAN: The Individual Aggregation Network for Person Search

  4. Neural Person Search Machines

Re-ID with GAN

  1. Unlabeled Samples Generated by GAN Improve the Person Re-identification Baseline in vitro
  2. Person Transfer GAN to Bridge Domain Gap for Person Re-Identification

Vehicle Re-ID

  1. Learning Deep Neural Networks for Vehicle Re-ID with Visual-spatio-temporal Path Proposals

Deep Metric Learning

  1. Deep Metric Learning for Person Re-Identification
  2. Deep Metric Learning for Practical Person Re-Identification
  3. Constrained Deep Metric Learning for Person Re-identification
  4. DarkRank: Accelerating Deep Metric Learning via Cross Sample Similarities Transfer

Re-ID with Attributes Prediction

  1. Deep Attributes Driven Multi-Camera Person Re-identification
  2. Improving Person Re-identification by Attribute and Identity Learning

Video-based Person Re-Identification

  1. Recurrent Convolutional Network for Video-based Person Re-Identification
  2. Deep Recurrent Convolutional Networks for Video-based Person Re-identification: An End-to-End Approach
  3. Jointly Attentive Spatial-Temporal Pooling Networks for Video-based Person Re-Identification
  4. Three-Stream Convolutional Networks for Video-based Person Re-Identification

Re-ranking

  1. Re-ranking Person Re-identification with k-reciprocal Encoding

实战项目

  1. Open-ReID: Open source person re-identification library in python
  2. caffe-PersonReID
  3. DukeMTMC-reID_baseline Matlab
  4. Code for IDE baseline on Market-1501

教程

  1. 1st Workshop on Target Re-Identification and Multi-Target Multi-Camera Tracking
  2. 郑哲东 -Deep-ReID:行人重识别的深度学习方法
  3. Person Identification in Large Scale Camera Networks Wei-Shi Zheng (郑伟诗)
  4. Person Re-Identification: Theory and Best Practice

数据集

  1. Re-ID 数据集汇总

图像数据集

  1. Market-1501 Dataset 751个人,27种属性,一共约三万张图像(一人多图)
  2.  DukeMTMC-reID DukeMTMC数据集的行人重识别子集,原始数据集地址(http://vision.cs.duke.edu/DukeMTMC/) ,为行人跟踪数据集。原始数据集包含了85分钟的高分辨率视频,采集自8个不同的摄像头。并且提供了人工标注的bounding box。最终,DukeMTMC-reID 包含了 16,522张训练图片(来自702个人), 2,228个查询图像(来自另外的702个人),以及 17,661 张图像的搜索库(gallery)。并提供切割后的图像供下载。
  3. CUHK01, 02, 03

Attribute相关数据集

  1. RAP
  2. Attribute for Market-1501and DukeMTMC_reID

视频相关数据集

  1. Mars
  2. PRID2011

NLP相关数据集:

  1. 自然语言搜图像
  2. 自然语言搜索行人所在视频

领域专家

  1. Shaogang Gong -[http://www.eecs.qmul.ac.uk/~sgg/]
  2. Xiaogang Wang
  3. Weishi Zheng
  4. Liang Zheng
  5. Chen Change Loy
  6. Qi Tian
  7. Shengcai Liao
  8. Rui Zhao
  9. Yang Yang
  10. Ling Shao
  11. Ziyan Wu
  12. DaPeng Chen
  13. Horst Bischof
  14. Niki Martinel
  15. Liang Lin
  16. Le An
  17. Xiang Bai
  18. Xiaoyuan Jing
  19. Fei Xiong
  20. DaPeng Chen

初步版本,水平有限,有错误或者不完善的地方,欢迎大家提建议和补充,会一直保持更新,本文为专知内容组原创内容,未经允许不得转载,如需转载请发送邮件至fangquanyi@gmail.com 或 联系微信专知小助手(Rancho_Fang)

敬请关注http://www.zhuanzhi.ai 和关注专知公众号,获取第一手AI相关知识

VIP内容

摘要: 行人重识别是近年来计算机视觉领域的热点问题, 经过多年的发展, 基于可见光图像的一般行人重识别技术已经趋近成熟. 然而, 目前的研究多基于一个相对理想的假设, 即行人图像都是在光照充足的条件下拍摄的高分辨率图像. 因此虽然大多数的研究都能取得较为满意的效果, 但在实际环境中并不适用. 多源数据行人重识别即利用多种行人信息进行行人匹配的问题. 除了需要解决一般行人重识别所面临的问题外, 多源数据行人重识别技术还需要解决不同类型行人信息与一般行人图片相互匹配时的差异问题, 如低分辨率图像、红外图像、深度图像、文本信息和素描图像等. 因此, 与一般行人重识别方法相比, 多源数据行人重识别研究更具实用性, 同时也更具有挑战性. 本文首先介绍了一般行人重识别的发展现状和所面临的问题, 然后比较了多源数据行人重识别与一般行人重识别的区别, 并根据不同数据类型总结了5 类多源数据行人重识别问题, 分别从方法、数据集两个方面对现有工作做了归纳和分析. 与一般行人重识别技术相比, 多源数据行人重识别的优点是可以充分利用各类数据学习跨模态和类型的特征转换. 最后, 本文讨论了多源数据行人重识别未来的发展.

http://www.aas.net.cn/cn/article/doi/10.16383/j.aas.c190278

成为VIP会员查看完整内容
0
15

最新论文

Authentication is the task of confirming the matching relationship between a data instance and a given identity. Typical examples of authentication problems include face recognition and person re-identification. Data-driven authentication could be affected by undesired biases, i.e., the models are often trained in one domain (e.g., for people wearing spring outfits) while applied in other domains (e.g., they change the clothes to summer outfits). Previous works have made efforts to eliminate domain-difference. They typically assume domain annotations are provided, and all the domains share classes. However, for authentication, there could be a large number of domains shared by different identities/classes, and it is impossible to annotate these domains exhaustively. It could make domain-difference challenging to model and eliminate. In this paper, we propose a domain-agnostic method that eliminates domain-difference without domain labels. We alternately perform latent domain discovery and domain-difference elimination until our model no longer detects domain-difference. In our approach, the latent domains are discovered by learning the heterogeneous predictive relationships between inputs and outputs. Then domain-difference is eliminated in both class-dependent and class-independent components. Comprehensive empirical evaluation results are provided to demonstrate the effectiveness and superiority of our proposed method.

0
0
下载
预览
Top