Abstract
This paper aims at studying unsupervised person re-identification (re-ID) which does not require any annotations. Recently, many approaches tackle this problem through contrastive learning due to its effective feature representation for unsupervised tasks. Especially, a uni-centroid representation is always obtained by averaging all the instance features within a cluster having the same pseudolabel. However, due to the unsatisfied clustering results, a cluster often contains some noisy samples, making the generated centroids imperfect. To address this issue, we propose a new graph correlation module (GCM) that can adaptively mine the relationship between each sample within the cluster and a high-quality relation-aware centroid is formed for momentum updating. Moreover, to increase the complexity of the task and prevent the model from falling into a local optimum, the original features extracted from the model are directly used to update the corresponding centroid. Extensive experiments demonstrate the superiority of the proposed method over state-of-the-art approaches on fully unsupervised re-ID tasks.
Similar content being viewed by others
Data availability
The datasets generated during the current study are available in the Market-1501 and DukeMTMC-re-ID.
References
Sun, Y., Zheng, L., Yang, Y., Tian, Q., Wang, S.: Beyond part models: Person retrieval with refined part pooling. In: Proceedings of the European Conference on Computer Vision (2018)
Luo, H., Gu, Y., Liao, X., Lai, S., Jiang, W.: Bag of tricks and a strong baseline for deep person re-identification. In: CVPR Workshops (2019)
Chen, H., Lagadec, B., Bremond, F.: Learning discriminative and generalizable representations by spatial-channel partition for person re-identification. In: WACV (2020)
Wu, Q., Dai, P., Chen, P., et al.: Deep adversarial data augmentation with attribute guided for person re-identification. In: Proceedings of the SIViP 15, 655–662 (2021)
Zou, Y., Yang, X., Yu, Z., Kumar, B.V.K., Kautz, J.: Joint disentangling and adaptation for crossdomain person re-identification. In: Proceedings of the European Conference on Computer Vision (2020)
Zhong, Z., Zheng, L., Luo, Z., Li, S., Yang, Y.: Invariance matters: exemplar memory for domain adaptive person re-identification. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 598–607 (2019)
Lin, Y., Dong, X., Zheng, L., Yan, Y., Yang, Y.: A bottom-up clustering approach to unsupervised person re-identification. Proc. AAAI Conf. Artif. Intell. 33, 8738–8745 (2019)
Lin, Y., Xie, L., Wu, Y., Yan, C., Tian, Q.: Unsupervised person re-identification via softened similarity learning. In: CVPR, pp. 3390–3399 (2020)
Zeng, K., Ning, M., Wang, Y., Guo, Y.: Hierarchical clustering with hard-batch triplet loss for person re-identification. In: CVPR, pp. 657–665 (2020)
Ge, Y., Zhu, F., Chen, D., Zhao, R.: Self-paced contrastive learning with hybrid memory for domain adaptive object re-id. Adv. Neural Inf. Process. Syst., 33, 11309-11321 (2020)
Dai, Z., Wang, G., Yuan, W., et al.: Cluster contrast for unsupervised person re-identification. arXiv:2103.11568 (2021)
Chen, H., Lagadec, B., Bremond, F.: ICE: inter-instance contrastive encoding for unsupervised person re-identification. CoRR, arXiv:2103.16364 (2021)
Wang, D., Zhang, S.: Unsupervised person re-identification via multi-label classification. In: CVPR (2020)
Wang, M., Lai, B., Huang, J., Gong, X., Hua, X.-S.: Camera-aware proxies for unsupervised person re-identification. In: AAAI (2021)
Wu, Y., Wu, X., Li, X., Tian, J.: MGH:metadata guided hypergraph modeling for unsupervised person re-identification. In: ACM Multimedia, pp.1571–1580 (2021)
Chen, T., Kornblith, S., Norouzi, M., et al.: A simple framework for contrastive learning of visual representations. In: International Conference on Machine Learning, PMLR, pp. 1597–1607 (2020)
He, K., Fan, H., Wu, Y., Xie, S., Girshick, R.: Momentum contrast for unsupervised visual representation learning. In: CVPR, pp. 9729–9738 (2020)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. In: CVPR, pp. 770–778 (2016)
MacQueen, J., et al.: Some methods for classification and analysis of multivariate observations. In: Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability 1, 281–297 (1967)
Ester, M., Kriegel, H.-P., Sander, J., Xiaowei, X., et al.: A density-based algorithm for discovering clusters in large spatial databases with noise. ACM SIGKDD Conf. Knowl. Discov. Data Mining 96, 226–231 (1996)
Li, S., Bak, S., Carr, P., Wang, X.: Diversity regularized spatiotemporal attention for video-based person reidentification. In: CVPR, 369–378 (2018)
Wang, X., Girshick, R., Gupta, A., He, K.: Non-local neural networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 7794–7803 (2018)
Zhang, X., Hou, M., Deng, X. et al.: Multi-cascaded attention and overlapping part features network for person re-identification. In: Proceedings of the SIViP (2022)
van den Oord, A., Li, Y., Vinyals, O.: Representation learning with contrastive predictive coding. arXiv:1807.03748, (2018)
Zheng, L., Shen, L., Tian, L., Wang, S., Wang, J., Tian, Q.: Scalable person re-identification: A benchmark. In: ICCV, pp. 1116–1124 (2015)
Ristani, E., Solera, F., Zou, R., Cucchiara, R., Tomasi, C.: Performance measures and a data set for multi-target, multi-camera tracking. In: ECCV, pp. 17–35 (2016)
Zhang, Z.., Lan, C., Zeng, W., Jin, X., Chen, Z.: Relation-aware global attention for person re-identification. In: CVPR, pp. 3186–3195 (2020)
Fu, Y., Wei, Y., Wang, G., Zhou, Y., Shi, H., Huang, T.S.: Self-similarity grouping: A simple unsupervised cross domain adaptation approach for person re-identification. In: ICCV, 6112–6121 (2019)
Zhai, Y., Lu, S., Ye, Q., Shan, X., Tian, Y.: Ad-cluster: Augmented discriminative clustering for domain adaptive person re-identification. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2020)
Ge, Y., Chen, D., Li, H.: Mutual mean-teaching: Pseudo label refinery for unsupervised domain adaptation on person re-identification. In: International Conference on Learning Representations (2020)
Liu, X., Zhang, S.: Graph consistency based mean-teaching for unsupervised domain adaptive person re-identification. In: IJCAI, 874–880 (2021)
Zheng, K., Lan, C., Zeng, W., Zhang, Z., Zha, Z.-J.: Exploiting sample uncertainty for domain adaptive person re-identification. In: AAAI, 3538–3546 (2021)
Zheng, K., Liu, W., He, L., Mei, T., Luo, J., Zha, Z.-J.: Group-aware label transfer for domain adaptive person re-identification. In: CVPR, 5310–5319 (2021)
Zheng, Y., Tang, S., Teng, G., Ge, Y., Liu, K., Qin, J., Qi, D., Chen, D.: Online pseudo label generation by hierarchical cluster dynamics for adaptive person re-identification. In: ICCV, pp. 8371–8381 (2021)
Wang, Z., Zhang, J., Zheng, L., Liu, Y., Sun, Y., Li, Y., Wang, S.: Cycas: Self-supervised cycle association for learning re-identifiable descriptions. In: CycAs: Self-supervised Cycle Association for Learning Re-identifiable Descriptions (2020)
Xuan, S., Zhang, S.: Intra-inter camera similarity for unsupervised person re-identification. In: CVPR, 11926–11935 ( 2021)
Acknowledgements
This research is partly supported by NSFC (62176169), and SCU-Luzhou Municipal Peoples Government Strategic Cooperation Project (2020CDLZ-10).
Funding
Not applicable.
Author information
Authors and Affiliations
Contributions
Xin Zhang contributed to the conception of the study and performed the experiment; Xin Zhang performed the data analyses and wrote the manuscript; Keren Fu contributed significantly to analysis and manuscript revision; Yanci Zhang helped perform the analysis with constructive discussions. All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Ethics approval and consent to participate
Not applicable.
Consent for publication
Not applicable.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Zhang, X., Fu, K. & Zhang, Y. Graph correlation-refined centroids for unsupervised person re-identification. SIViP 17, 1457–1464 (2023). https://doi.org/10.1007/s11760-022-02354-5
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11760-022-02354-5