Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

A Discriminatively Learned CNN Embedding for Person Reidentification

A Discriminatively Learned CNN Embedding for Person Reidentification In this article, we revisit two popular convolutional neural networks in person re-identification (re-ID): verification and identification models. The two models have their respective advantages and limitations due to different loss functions. Here, we shed light on how to combine the two models to learn more discriminative pedestrian descriptors. Specifically, we propose a Siamese network that simultaneously computes the identification loss and verification loss. Given a pair of training images, the network predicts the identities of the two input images and whether they belong to the same identity. Our network learns a discriminative embedding and a similarity measurement at the same time, thus taking full usage of the re-ID annotations. Our method can be easily applied on different pretrained networks. Albeit simple, the learned embedding improves the state-of-the-art performance on two public person re-ID benchmarks. Further, we show that our architecture can also be applied to image retrieval. The code is available at https://github.com/layumi/2016_person_re-ID. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM) Association for Computing Machinery

A Discriminatively Learned CNN Embedding for Person Reidentification

Download PDF
20 pages

Loading next page...
 
/lp/association-for-computing-machinery/a-discriminatively-learned-cnn-embedding-for-person-reidentification-nBB1VIH7bQ
Publisher
Association for Computing Machinery
Copyright
Copyright © 2017 ACM
ISSN
1551-6857
eISSN
1551-6865
DOI
10.1145/3159171
Publisher site
See Article on Publisher Site

Abstract

In this article, we revisit two popular convolutional neural networks in person re-identification (re-ID): verification and identification models. The two models have their respective advantages and limitations due to different loss functions. Here, we shed light on how to combine the two models to learn more discriminative pedestrian descriptors. Specifically, we propose a Siamese network that simultaneously computes the identification loss and verification loss. Given a pair of training images, the network predicts the identities of the two input images and whether they belong to the same identity. Our network learns a discriminative embedding and a similarity measurement at the same time, thus taking full usage of the re-ID annotations. Our method can be easily applied on different pretrained networks. Albeit simple, the learned embedding improves the state-of-the-art performance on two public person re-ID benchmarks. Further, we show that our architecture can also be applied to image retrieval. The code is available at https://github.com/layumi/2016_person_re-ID.

Journal

ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM) – Association for Computing Machinery

Published: Dec 13, 2017

Keywords: Person reidentification

References

  • An improved deep learning architecture for person re-identification
    Ahmed, Ejaz; Jones, Michael; Marks, Tim K.
  • Neural codes for image retrieval
    Babenko, Artem; Slesarev, Anton; Chigorin, Alexandr; Lempitsky, Victor
  • Signature verification using a Siamese time delay neural network
    Bromley, Jane; Bentz, James W.; Bottou, Léon; Guyon, Isabelle; LeCun, Yann; Moore, Cliff; Säckinger, Eduard; Shah, Roopak
  • Semisupervised feature analysis by mining correlations among multiple tasks
    Chang, Xiaojun; Yang, Yi
  • Semantic pooling for complex event analysis in untrimmed videos
    Chang, Xiaojun; Yu, Yao-Liang; Yang, Yi; Xing, Eric P.
  • Similarity learning with spatial constraints for person re-identification
    Chen, Dapeng; Yuan, Zejian; Chen, Badong; Zheng, Nanning
  • Person re-identification by multi-channel parts-based CNN with improved triplet loss function
    Cheng, De; Gong, Yihong; Zhou, Sanping; Wang, Jinjun; Zheng, Nanning
  • Discriminative dictionary learning with common label alignment for cross-modal retrieval
    Deng, Cheng; Tang, Xu; Yan, Junchi; Liu, Wei; Gao, Xinbo
  • Deep feature learning with relative distance comparison for person re-identification
    Ding, Shengyong; Lin, Liang; Wang, Guangrun; Chao, Hongyang
  • Deep transfer learning for person re-identification
    Geng, Mengyue; Wang, Yaowei; Xiang, Tao; Tian, Yonghong
  • Dimensionality reduction by learning an invariant mapping
    Hadsell, Raia; Chopra, Sumit; LeCun, Yann
  • Deep residual learning for image recognition
    He, Kaiming; Zhang, Xiangyu; Ren, Shaoqing; Sun, Jian
  • Query expansion for content-based similarity search using local and global features
    Houle, Michael E.; Ma, Xiguo; Oria, Vincent; Sun, Jichao
  • Cross-dimensional weighting for aggregated deep convolutional features
    Kalantidis, Yannis; Mellina, Clayton; Osindero, Simon
  • Large scale metric learning from equivalence constraints
    Koestinger, Martin; Hirzer, Martin; Wohlhart, Paul; Roth, Peter M.; Bischof, Horst
  • Large scale metric learning from equivalence constraints
    Köstinger, Martin; Hirzer, Martin; Wohlhart, Paul; Roth, Peter M.; Bischof, Horst
  • ImageNet classification with deep convolutional neural networks
    Krizhevsky, Alex; Sutskever, Ilya; Hinton, Geoffrey E.
  • DeepReID: Deep filter pairing neural network for person re-identification
    Li, Wei; Zhao, Rui; Xiao, Tong; Wang, Xiaogang
  • Person re-identification by local maximal occurrence representation and metric learning
    Liao, Shengcai; Hu, Yang; Zhu, Xiangyu; Li, Stan Z.
  • Multichannel-kernel canonical correlation analysis for cross-view person reidentification
    Lisanti, Giuseppe; Karaman, Svebor; Masi, Iacopo
  • End-to-end comparative attention networks for person re-identification
    Liu, Hao; Feng, Jiashi; Qi, Meibin; Jiang, Jianguo; Yan, Shuicheng
  • Large-margin softmax loss for convolutional neural networks
    Liu, Weiyang; Wen, Yandong; Yu, Zhiding; Yang, Meng
  • Hierarchical Gaussian descriptor for person re-identification
    Matsukawa, Tetsu; Okabe, Takahiro; Suzuki, Einoshin; Sato, Yoichi
  • PCCA: A new approach for distance learning from sparse pairwise constraints
    Mignon, Alexis; Jurie, Frédéric
  • Multi-camera coordination and control in surveillance systems: A survey
    Natarajan, Prabhu; Atrey, Pradeep K.; Kankanhalli, Mohan
  • Deep metric learning via lifted structured feature embedding
    Song, Hyun Oh; Xiang, Yu; Jegelka, Stefanie; Savarese, Silvio
  • Object retrieval with large vocabularies and fast spatial matching
    Philbin, James; Chum, Ondrej; Isard, Michael; Sivic, Josef; Zisserman, Andrew
  • Multiple measurements and joint dimensionality reduction for large scale image search with short vectors
    Radenović, Filip; Jégou, Hervé; Chum, Ondrej
  • CNN image retrieval learns from BoW: Unsupervised fine-tuning with hard examples
    Radenović, Filip; Tolias, Giorgos; Chum, Ondřej
  • ImageNet large scale visual recognition challenge
    Russakovsky, Olga; Deng, Jia; Su, Hao; Krause, Jonathan; Satheesh, Sanjeev; Ma, Sean; Huang, Zhiheng
  • Very deep convolutional networks for large-scale image recognition
    Simonyan, Karen; Zisserman, Andrew
  • Dropout: A simple way to prevent neural networks from overfitting
    Srivastava, Nitish; Hinton, Geoffrey E.; Krizhevsky, Alex; Sutskever, Ilya; Salakhutdinov, Ruslan
  • Deep learning face representation by joint identification-verification
    Sun, Yi; Chen, Yuheng; Wang, Xiaogang; Tang, Xiaoou
  • Deeply learned face representations are sparse, selective, and robust
    Sun, Yi; Wang, Xiaogang; Tang, Xiaoou
  • Particular object retrieval with integral max-pooling of CNN activations
    Tolias, Giorgos; Sicre, Ronan; Jégou, Hervé
  • Multiregion bilinear convolutional neural networks for person re-identification
    Ustinova, Evgeniya; Ganin, Yaroslav; Lempitsky, Victor
  • Accelerating t-SNE using tree-based algorithms
    Van Der Maaten, Laurens
  • Gated Siamese convolutional neural network architecture for human re-identification
    Varior, Rahul Rama; Haloi, Mrinal; Wang, Gang
  • A Siamese long short-term memory architecture for human re-identification
    Varior, Rahul Rama; Shuai, Bing; Lu, Jiwen; Xu, Dong; Wang, Gang
  • MatConvNet—convolutional neural networks for MATLAB
    Vedaldi, A.; Lenc, K.
  • Zero-shot person re-identification via cross-view consistency
    Wang, Zheng; Hu, Ruimin; Liang, Chao; Yu, Yi; Jiang, Junjun; Ye, Mang; Chen, Jun; Leng, Qingming
  • Deep linear discriminant analysis on Fisher networks: A hybrid architecture for person re-identification
    Wu, Lin; Shen, Chunhua; van den Hengel, Anton
  • PersonNet: Person re-identification with deep convolutional neural networks
    Wu, Lin; Shen, Chunhua; van den Hengel, Anton
  • Image classification by cross-media active learning with privileged information
    Yan, Yan; Nie, Feiping; Li, Wen; Gao, Chenqiang; Yang, Yi; Xu, Dong
  • Enhancing person re-identification in a self-trained subspace
    Yang, Xun; Wang, Meng; Hong, Richang; Tian, Qi; Rui, Yong
  • Ranking with local regression and global alignment for cross media retrieval
    Yang, Yi; Xu, Dong; Nie, Feiping; Luo, Jiebo; Zhuang, Yueting
  • Deep metric learning for person re-identification
    Yi, Dong; Lei, Zhen; Liao, Shengcai; Li, Stan Z.
  • Learning a discriminative null space for person re-identification
    Zhang, Li; Xiang, Tao; Gong, Shaogang
  • Sample-specific SVM learning for person re-identification
    Zhang, Ying; Li, Baohua; Lu, Huchuan; Irie, Atshushi; Ruan, Xiang
  • Person re-identification by salience matching
    Zhao, Rui; Ouyang, Wanli; Wang, Xiaogang
  • MARS: A video benchmark for large-scale person re-identification
    Zheng, Liang; Bie, Zhi; Sun, Yifan; Wang, Jingdong; Su, Chi; Wang, Shengjin; Tian, Qi
  • Scalable person re-identification: A benchmark
    Zheng, Liang; Shen, Liyue; Tian, Lu; Wang, Shengjin; Wang, Jingdong; Tian, Qi
  • Fast image retrieval: Query pruning and early termination
    Zheng, Liang; Wang, Shengjin; Liu, Ziqiong; Tian, Qi
  • Coupled binary embedding for large-scale image retrieval
    Zheng, Liang; Wang, Shengjin; Tian, Qi
  • Accurate image search with multi-scale contextual evidences
    Zheng, Liang; Wang, Shengjin; Wang, Jingdong; Tian, Qi
  • Person re-identification: Past, present and future
    Zheng, Liang; Yang, Yi; Hauptmann, Alexander G.
  • SIFT meets CNN: A decade survey of instance retrieval
    Zheng, Liang; Yang, Yi; Tian, Qi
  • Reidentification by relative distance comparison
    Zheng, Wei-Shi; Gong, Shaogang; Xiang, Tao
  • Unlabeled samples generated by GAN improve the person re-identification baseline in vitro
    Zheng, Zhedong; Zheng, Liang; Yang, Yi