Abstract
Due to domain bias, directly deploying a deep person re-identification (re-ID) model trained on one dataset often achieves considerably poor accuracy on another dataset. In this article, we propose an Adaptive Exploration (AE) method to address the domain-shift problem for re-ID in an unsupervised manner. Specifically, in the target domain, the re-ID model is inducted to (1) maximize distances between all person images and (2) minimize distances between similar person images. In the first case, by treating each person image as an individual class, a non-parametric classifier with a feature memory is exploited to encourage person images to move far away from each other. In the second case, according to a similarity threshold, our method adaptively selects neighborhoods for each person image in the feature space. By treating these similar person images as the same class, the non-parametric classifier forces them to stay closer. However, a problem of the adaptive selection is that, when an image has too many neighborhoods, it is more likely to attract other images as its neighborhoods. As a result, a minority of images may select a large number of neighborhoods while a majority of images has only a few neighborhoods. To address this issue, we additionally integrate a balance strategy into the adaptive selection. We evaluate our methods with two protocols. The first one is called “target-only re-ID”, in which only the unlabeled target data is used for training. The second one is called “domain adaptive re-ID”, in which both the source data and the target data are used during training. Experimental results on large-scale re-ID datasets demonstrate the effectiveness of our method. Our code has been released at https://github.com/dyh127/Adaptive-Exploration-for-Unsupervised-Person-Re-Identification.
- Shi-Zhe Chen, Chun-Chao Guo, and Jian-Huang Lai. 2016. Deep ranking for person re-identification via joint representation learning. TIP 25, 5 (2016), 2353--2367. DOI:https://doi.org/10.1109/TIP.2016.2545929Google Scholar
Cross Ref
- Weihua Chen, Xiaotang Chen, Jianguo Zhang, and Kaiqi Huang. 2017. Beyond triplet loss: A deep quadruplet network for person re-identification. In CVPR. 1320--1329. DOI:https://doi.org/10.1109/CVPR.2017.145Google Scholar
- De Cheng, Yihong Gong, Sanping Zhou, Jinjun Wang, and Nanning Zheng. 2016. Person re-identification by multi-channel parts-based CNN with improved triplet loss function. In CVPR. 1335--1344. DOI:https://doi.org/10.1109/CVPR.2016.149Google Scholar
- Yeong-Jun Cho and Kuk-Jin Yoon. 2016. Improving person re-identification via pose-aware multi-shot matching. In CVPR. 1354--1362. DOI:https://doi.org/10.1109/CVPR.2016.151Google Scholar
- Cheng Deng, Xianglong Liu, Chao Li, and Dacheng Tao. 2018. Active multi-kernel domain adaptation for hyperspectral image classification. PR 77 (2018), 306--315. DOI:https://doi.org/10.1016/j.patcog.2017.10.007Google Scholar
- Cheng Deng, Yumeng Xue, Xianglong Liu, Chao Li, and Dacheng Tao. 2019. Active transfer learning network: A unified deep joint spectral-spatial feature learning model for hyperspectral image classification. TGRS 57, 3 (2019), 1741--1754. DOI:https://doi.org/10.1109/TGRS.2018.2868851Google Scholar
- Jia Deng, Wei Dong, Richard Socher, Li-Jia Li, Kai Li, and Fei-Fei Li. 2009. ImageNet: A large-scale hierarchical image database. In CVPR. 248--255. DOI:https://doi.org/10.1109/CVPRW.2009.5206848Google Scholar
- Weijian Deng, Liang Zheng, Qixiang Ye, Guoliang Kang, Yi Yang, and Jianbin Jiao. 2018. Image-image domain adaptation with preserved self-similarity and domain-dissimilarity for person re-identification. In CVPR. 994--1003. DOI:https://doi.org/10.1109/CVPR.2018.00110Google Scholar
- Hehe Fan, Xiaojun Chang, De Cheng, Yi Yang, Dong Xu, and Alexander G. Hauptmann. 2017. Complex event detection by identifying reliable shots from untrimmed videos. In ICCV. 736--744. DOI:https://doi.org/10.1109/ICCV.2017.86Google Scholar
- Hehe Fan, Liang Zheng, Chenggang Yan, and Yi Yang. 2018. Unsupervised person re-identification: Clustering and fine-tuning. TOMCCAP 14, 4 (2018), 83:1--83:18. DOI:https://doi.org/10.1145/3243316Google Scholar
- Michela Farenzena, Loris Bazzani, Alessandro Perina, Vittorio Murino, and Marco Cristani. 2010. Person re-identification by symmetry-driven accumulation of local features. In CVPR. 2360--2367. DOI:https://doi.org/10.1109/CVPR.2010.5539926Google Scholar
- Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2016. Deep residual learning for image recognition. In CVPR. 770--778. DOI:https://doi.org/10.1109/CVPR.2016.90Google Scholar
- Alexander Hermans, Lucas Beyer, and Bastian Leibe. 2017. In defense of the triplet loss for person re-identification. arXiv:1703.07737 (2017).Google Scholar
- Sergey Ioffe and Christian Szegedy. 2015. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In ICML. 448--456.Google Scholar
- Srikrishna Karanam, Yang Li, and Richard J. Radke. 2015. Person re-identification with discriminatively trained viewpoint invariant dictionaries. In ICCV. 4516--4524. DOI:https://doi.org/10.1109/ICCV.2015.513Google Scholar
- Martin Köstinger, Martin Hirzer, Paul Wohlhart, Peter M. Roth, and Horst Bischof. 2012. Large scale metric learning from equivalence constraints. In CVPR. 2288--2295. DOI:https://doi.org/10.1109/CVPR.2012.6247939Google Scholar
- M. Pawan Kumar, Benjamin Packer, and Daphne Koller. 2010. Self-paced learning for latent variable models. In NeurIPS. 1189--1197.Google Scholar
- Wei Li, Rui Zhao, Tong Xiao, and Xiaogang Wang. 2014. DeepReID: Deep filter pairing neural network for person re-identification. In CVPR. 152--159. DOI:https://doi.org/10.1109/CVPR.2014.27Google Scholar
- Yu-Jhe Li, Fu-En Yang, Yen-Cheng Liu, Yu-Ying Yeh, Xiaofei Du, and Yu-Chiang Frank Wang. 2018. Adaptation and re-identification network: An unsupervised deep transfer learning approach to person re-identification. In CVPR. 172--178. DOI:https://doi.org/10.1109/CVPRW.2018.00054Google Scholar
- Shengcai Liao, Yang Hu, Xiangyu Zhu, and Stan Z. Li. 2015. Person re-identification by local maximal occurrence representation and metric learning. In CVPR. 2197--2206. DOI:https://doi.org/10.1109/CVPR.2015.7298832Google Scholar
- Shan Lin, Haoliang Li, Chang-Tsun Li, and Alex C. Kot. 2018. Multi-task mid-level feature alignment network for unsupervised cross-dataset person re-identification. In BMVC. 9.Google Scholar
- Yutian Lin, Xuanyi Dong, Liang Zheng, Yan Yan, and Yi Yang. 2019. A bottom-up clustering approach to unsupervised person re-identification. In AAAI. 8738--8745. DOI:https://doi.org/10.1609/aaai.v33i01.33018738Google Scholar
- Hao Liu, Jiashi Feng, Meibin Qi, Jianguo Jiang, and Shuicheng Yan. 2017. End-to-end comparative attention networks for person re-identification. TIP 26, 7 (2017), 3492--3506. DOI:https://doi.org/10.1109/TIP.2017.2700762Google Scholar
Cross Ref
- Zimo Liu, Dong Wang, and Huchuan Lu. 2017. Stepwise metric promotion for unsupervised video person re-identification. In ICCV. 2448--2457. DOI:https://doi.org/10.1109/ICCV.2017.266Google Scholar
- Fan Ma, Deyu Meng, Qi Xie, Zina Li, and Xuanyi Dong. 2017. Self-paced co-training. In ICML. 2275--2284.Google Scholar
- Laurens van der Maaten and Geoffrey Hinton. 2008. Visualizing data using t-SNE. JMLR 9, Nov (2008), 2579--2605.Google Scholar
- Jiaxu Miao, Yu Wu, Ping Liu, Yuhang Ding, and Yi Yang. 2019. Pose-guided feature alignment for occluded person re-identification. ICCV (2019).Google Scholar
- Vinod Nair and Geoffrey E. Hinton. 2010. Rectified linear units improve restricted Boltzmann machines. In ICML. 807--814.Google Scholar
- Ruijie Quan, Xuanyi Dong, Yu Wu, Linchao Zhu, and Yi Yang. 2019. Auto-ReID: Searching for a part-aware ConvNet for person re-identification. ICCV (2019).Google Scholar
- M. Saquib Sarfraz, Arne Schumann, Andreas Eberle, and Rainer Stiefelhagen. 2018. A pose-sensitive embedding for person re-identification with expanded cross neighborhood re-ranking. In CVPR. 420--429. DOI:https://doi.org/10.1109/CVPR.2018.00051Google Scholar
- Hailin Shi, Yang Yang, Xiangyu Zhu, Shengcai Liao, Zhen Lei, Wei-Shi Zheng, and Stan Z. Li. 2016. Embedding deep metric for person re-identification: A study against large variations. In ECCV. 732--748. DOI:https://doi.org/10.1007/978-3-319-46448-0_44Google Scholar
- Chunfeng Song, Yan Huang, Wanli Ouyang, and Liang Wang. 2018. Mask-guided contrastive attention model for person re-identification. In CVPR. 1179--1188. DOI:https://doi.org/10.1109/CVPR.2018.00129Google Scholar
- Jingkuan Song, Yuyu Guo, Lianli Gao, Xuelong Li, Alan Hanjalic, and Heng Tao Shen. 2018. From deterministic to generative: Multi-modal stochastic RNNs for video captioning. TNNLS 30, 10 (2018), 3047--3058. DOI:https://doi.org/10.1109/TNNLS.2018.2851077Google Scholar
Cross Ref
- Jingkuan Song, Hanwang Zhang, Xiangpeng Li, Lianli Gao, Meng Wang, and Richang Hong. 2018. Self-supervised video hashing with hierarchical binary auto-encoder. TIP 27, 7 (2018), 3210--3221. DOI:https://doi.org/10.1109/TIP.2018.2814344Google Scholar
Cross Ref
- Nitish Srivastava, Geoffrey E. Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. Dropout: A simple way to prevent neural networks from overfitting. JMLR 15, 1 (2014), 1929--1958.Google Scholar
Digital Library
- Chi Su, Jianing Li, Shiliang Zhang, Junliang Xing, Wen Gao, and Qi Tian. 2017. Pose-driven deep convolutional model for person re-identification. In ICCV. 3980--3989. DOI:https://doi.org/10.1109/ICCV.2017.427Google Scholar
- Xiaoxiao Sun and Liang Zheng. 2019. Dissecting person re-identification from the viewpoint of viewpoint. CVPR (2019), 608--617.Google Scholar
- Maoqing Tian, Shuai Yi, Hongsheng Li, Shihua Li, Xuesen Zhang, Jianping Shi, Junjie Yan, and Xiaogang Wang. 2018. Eliminating background-bias for robust person re-identification. In CVPR. 5794--5803. DOI:https://doi.org/10.1109/CVPR.2018.00607Google Scholar
- Rahul Rama Varior, Mrinal Haloi, and Gang Wang. 2016. Gated Siamese convolutional neural network architecture for human re-identification. In ECCV. 791--808. DOI:https://doi.org/10.1007/978-3-319-46484-8_48Google Scholar
- Rahul Rama Varior, Bing Shuai, Jiwen Lu, Dong Xu, and Gang Wang. 2016. A Siamese long short-term memory architecture for human re-identification. In ECCV. 135--153. DOI:https://doi.org/10.1007/978-3-319-46478-7_9Google Scholar
- Faqiang Wang, Wangmeng Zuo, Liang Lin, David Zhang, and Lei Zhang. 2016. Joint learning of single-image and cross-image representations for person re-identification. In CVPR. 1288--1296. DOI:https://doi.org/10.1109/CVPR.2016.144Google Scholar
- Jin Wang, Zheng Wang, Chao Liang, Changxin Gao, and Nong Sang. 2018. Equidistance constrained metric learning for person re-identification. PR 74 (2018), 38--51. DOI:https://doi.org/10.1016/j.patcog.2017.09.014Google Scholar
- Jingya Wang, Xiatian Zhu, Shaogang Gong, and Wei Li. 2018. Transferable joint attribute-identity deep learning for unsupervised person re-identification. In CVPR. 2275--2284. DOI:https://doi.org/10.1109/CVPR.2018.00242Google Scholar
- Longhui Wei, Shiliang Zhang, Wen Gao, and Qi Tian. 2018. Person transfer GAN to bridge domain gap for person re-identification. In CVPR. 79--88. DOI:https://doi.org/10.1109/CVPR.2018.00016Google Scholar
- Yu Wu, Yutian Lin, Xuanyi Dong, Yan Yan, Wei Bian, and Yi Yang. 2019. Progressive learning for person re-identification with one example. TIP 28, 6 (2019), 2872--2881. DOI:https://doi.org/10.1109/TIP.2019.2891895Google Scholar
Cross Ref
- Zhirong Wu, Alexei A. Efros, and Stella X. Yu. 2018. Improving generalization via scalable neighborhood component analysis. In ECCV. 712--728. DOI:https://doi.org/10.1007/978-3-030-01234-2_42Google Scholar
- Ziyan Wu, Yang Li, and Richard J. Radke. 2015. Viewpoint invariant human re-identification in camera networks using pose priors and subject-discriminative features. PAMI 37, 5 (2015), 1095--1108. DOI:https://doi.org/10.1109/TPAMI.2014.2360373Google Scholar
Cross Ref
- Zhirong Wu, Yuanjun Xiong, Stella X. Yu, and Dahua Lin. 2018. Unsupervised feature learning via non-parametric instance discrimination. In CVPR. 3733--3742. DOI:https://doi.org/10.1109/CVPR.2018.00393Google Scholar
- Tong Xiao, Shuang Li, Bochao Wang, Liang Lin, and Xiaogang Wang. 2017. Joint detection and identification feature learning for person search. In CVPR. 3376--3385. DOI:https://doi.org/10.1109/CVPR.2017.360Google Scholar
- Yan Yan, Feiping Nie, Wen Li, Chenqiang Gao, Yi Yang, and Dong Xu. 2016. Image classification by cross-media active learning with privileged information. TMM 18, 12 (2016), 2494--2502. DOI:https://doi.org/10.1109/TMM.2016.2602938Google Scholar
Digital Library
- Qize Yang, Hong-Xing Yu, Ancong Wu, and Wei-Shi Zheng. 2019. Patch-based discriminative feature learning for unsupervised person re-identification. In CVPR.Google Scholar
- Dong Yi, Zhen Lei, Shengcai Liao, and Stan Z. Li. 2014. Deep metric learning for person re-identification. In ICPR. 34--39. DOI:https://doi.org/10.1109/ICPR.2014.16Google Scholar
- Hong-Xing Yu, Ancong Wu, and Wei-Shi Zheng. 2017. Cross-view asymmetric metric learning for unsupervised person re-identification. In ICCV. 994--1002. DOI:https://doi.org/10.1109/ICCV.2017.113Google Scholar
- Rui Zhao, Wanli Ouyang, and Xiaogang Wang. 2013. Unsupervised salience learning for person re-identification. In CVPR. 3586--3593. DOI:https://doi.org/10.1109/CVPR.2013.460Google Scholar
- Kang Zheng, Xiaochuan Fan, Yuewei Lin, Hao Guo, Hongkai Yu, Dazhou Guo, and Song Wang. 2017. Learning view-invariant features for person identification in temporally synchronized videos taken by wearable cameras. In ICCV. 2877--2885. DOI:https://doi.org/10.1109/ICCV.2017.311Google Scholar
- Liang Zheng, Yujia Huang, Huchuan Lu, and Yi Yang. 2019. Pose invariant embedding for deep person re-identification. TIP 28, 9 (2019), 4500--4509. DOI:https://doi.org/10.1109/TIP.2019.2910414Google Scholar
Cross Ref
- Liang Zheng, Liyue Shen, Lu Tian, Shengjin Wang, Jingdong Wang, and Qi Tian. 2015. Scalable person re-identification: A benchmark. In ICCV. 1116--1124. DOI:https://doi.org/10.1109/ICCV.2015.133Google Scholar
- Liang Zheng, Hengheng Zhang, Shaoyan Sun, Manmohan Chandraker, Yi Yang, and Qi Tian. 2017. Person re-identification in the wild. In CVPR. 3346--3355. DOI:https://doi.org/10.1109/CVPR.2017.357Google Scholar
- Wei-Shi Zheng, Shaogang Gong, and Tao Xiang. 2013. Reidentification by relative distance comparison. TPAMI (2013), 653--668. DOI:https://doi.org/10.1109/TPAMI.2012.138Google Scholar
- Zhedong Zheng, Xiaodong Yang, Zhiding Yu, Liang Zheng, Yi Yang, and Jan Kautz. 2019. Joint discriminative and generative learning for person re-identification. CVPR (2019), 2138--2147.Google Scholar
- Zhedong Zheng, Liang Zheng, and Yi Yang. 2016. A discriminatively learned CNN embedding for person re-identification. TOMCCAP 14, 1 (2016), 13. DOI:https://doi.org/10.1145/3159171Google Scholar
- Zhedong Zheng, Liang Zheng, and Yi Yang. 2017. Unlabeled samples generated by GAN improve the person re-identification baseline in vitro. In ICCV. 3774--3782. DOI:https://doi.org/10.1109/ICCV.2017.405Google Scholar
- Zhun Zhong, Liang Zheng, Guoliang Kang, Shaozi Li, and Yi Yang. 2017. Random erasing data augmentation. arXiv:1708.04896 (2017).Google Scholar
- Zhun Zhong, Liang Zheng, Shaozi Li, and Yi Yang. 2018. Generalizing a person retrieval model hetero- and homogeneously. In ECCV. 176--192. DOI:https://doi.org/10.1007/978-3-030-01261-8_11Google Scholar
- Zhun Zhong, Liang Zheng, Zhiming Luo, Shaozi Li, and Yi Yang. 2019. Invariance matters: Exemplar memory for domain adaptive person re-identification. CVPR (2019), 598--607.Google Scholar
- Zhun Zhong, Liang Zheng, Zhedong Zheng, Shaozi Li, and Yi Yang. 2019. CamStyle: A novel data augmentation method for person re-identification. TIP 28, 3 (2019), 1176--1190. DOI:https://doi.org/10.1109/TIP.2018.2874313Google Scholar
Digital Library
- Jun-Yan Zhu, Taesung Park, Phillip Isola, and Alexei A. Efros. 2017. Unpaired image-to-image translation using cycle-consistent adversarial networks. In ICCV. 2242--2251. DOI:https://doi.org/10.1109/ICCV.2017.244Google Scholar
Index Terms
Adaptive Exploration for Unsupervised Person Re-identification
Recommendations
Momentum source-proxy guided initialization for unsupervised domain adaptive person re-identification
AbstractUnsupervised domain adaptive person re-identification (UDA Re-ID), aiming to adapt the model trained from source domain to target domain, is especially challenging due to the non-overlapping identities between the two Re-ID domains. ...
Unsupervised Person Re-Identification via Multi-Label Classification
AbstractThe challenge of unsupervised person re-identification (ReID) lies in learning discriminative features without true labels. Most of previous works predict single-class pseudo labels through clustering. To improve the quality of generated pseudo ...
Domain adaptation with structural knowledge transfer learning for person re-identification
AbstractRecently, supervised person re-identification (Re-ID) algorithms have achieved great performance on benchmarks. However, it highly depends on labeled training samples and may not generalize well to new domains, limiting the applicability of person ...






Comments