ABSTRACT
In the study of unsupervised domain adaptation for pedestrian re-identification, one prevalent approach during training on unlabeled target domains involves generating pseudo-labels. However, due to the inability to guarantee the correctness of pseudo-labels, they inevitably carry noise, subsequently affecting the network's training process. In this paper, we propose a novel approach that better constrains noise and demonstrates strong performance. First, building upon the DBSCAN clustering method, we generate soft labels for samples, thereby providing fine-grained category loss supervision. Specifically, we introduce an adaptive reverse cross-entropy loss to impose adaptive constraints on the pseudo-label generation process. We conduct various experiments on four datasets across three transfer tasks, and the results consistently highlight the superiority of our method, providing strong evidence for its effectiveness.
- Yifan Sun, Liang Zheng, Yi Yang, Qi Tian, and ShengjinWang. Beyond part models: Person retrieval with refined part pooling (and A strong convolutional baseline). In Computer Vision - ECCV 2018 - 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part IV,volume 11208, pages 501–518, 2018.111Google Scholar
- Guanshuo Wang, Yufeng Yuan, Xiong Chen, Jiwei Li, and Xi Zhou. Learning discriminative features with multiple granularities for person re-identification. In 2018 ACM Multimedia Conference on Multimedia Conference, MM 2018, Seoul,Republic of Korea, October 22-26, 2018, pages 274–282,2018.Google ScholarDigital Library
- Zhizheng Zhang, Cuiling Lan, Wenjun Zeng, Xin Jin, andZhibo Chen. Relation-aware global attention for person reidentification. In Proceedings of the IEEE/CVF Conferenceon Computer Vision and Pattern Recognition (CVPR), June2020.Google Scholar
- Bruna J, Zaremba W, Szlam A, Spectral networks and locally connected networks on graphs[J]. arXiv preprint arXiv:1312.6203, 2013.Google Scholar
- Kipf T N, Welling M. Semi-supervised classification with graph convolutional networks[J]. arXiv preprint arXiv:1609.02907, 2016.Google Scholar
- Zheng L, Shen L, Tian L, Scalable person re-identification: A benchmark[C]//Proceedings of the IEEE international conference on computer vision. 2015: 1116-1124.Google Scholar
- Ristani E, Solera F, Zou R, Performance measures and a data set for multi-target, multi-camera tracking[C]//Computer Vision–ECCV 2016 Workshops: Amsterdam, The Netherlands, October 8-10 and 15-16, 2016, Proceedings, Part II. Cham: Springer International Publishing, 2016: 17-35.Google Scholar
- Li W, Zhao R, Xiao T, Deepreid: Deep filter pairing neural network for person re-identification[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2014: 152-159.Google Scholar
- Wei L, Zhang S, Gao W, Person transfer gan to bridge domain gap for person re-identification[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 79-88.Google Scholar
- Xiang S, Fu Y, You G, Unsupervised domain adaptation through synthesis for person re-identification[C]//2020 IEEE International Conference on Multimedia and Expo (ICME). IEEE, 2020: 1-6.Google Scholar
- Goldberger J, Ben-Reuven E. Training deep neural-networks using a noise adaptation layer[C]//International conference on learning representations. 2017.Google Scholar
- Jiang L, Zhou Z, Leung T, Mentornet: Learning data-driven curriculum for very deep neural networks on corrupted labels[C]//International conference on machine learning. PMLR, 2018: 2304-2313.Google Scholar
- Shen Y, Sanghavi S. Learning with bad training data via iterative trimmed loss minimization[C]//International Conference on Machine Learning. PMLR, 2019: 5739-5748.Google Scholar
- Zhang Z, Sabuncu M. Generalized cross entropy loss for training deep neural networks with noisy labels[J]. Advances in neural information processing systems, 2018, 31.Google Scholar
- Zhong Z, Zheng L, Luo Z, Invariance matters: Exemplar memory for domain adaptive person re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2019: 598-607.Google Scholar
- Zhao F, Liao S, Xie G S, Unsupervised domain adaptation with noise resistible mutual-training for person re-identification[C]//Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part XI 16. Springer International Publishing, 2020: 526-544.Google Scholar
- Sun J, Li Y, Chen H, Unsupervised cross domain person re-identification by multi-loss optimization learning[J]. IEEE Transactions on Image Processing, 2021, 30: 2935-2946.Google ScholarDigital Library
- Jin X, Lan C, Zeng W, Global distance-distributions separation for unsupervised person re-identification[C]//Computer Vision–ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part VII 16. Springer International Publishing, 2020: 735-751.Google Scholar
- Yu H X, Zheng W S, Wu A, Unsupervised person re-identification by soft multilabel learning[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 2148-2157.Google Scholar
- Wang D, Zhang S. Unsupervised person re-identification via multi-label classification[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2020: 10981-10990.Google Scholar
- Fu Y, Wei Y, Wang G, Self-similarity grouping: A simple unsupervised cross domain adaptation approach for person re-identification[C]//proceedings of the IEEE/CVF international conference on computer vision. 2019: 6112-6121.Google Scholar
- Ge Y, Chen D, Li H. Mutual mean-teaching: Pseudo label refinery for unsupervised domain adaptation on person re-identification[J]. arXiv preprint arXiv:2001.01526, 2020.Google Scholar
- Ge Y, Zhu F, Chen D, Self-paced contrastive learning with hybrid memory for domain adaptive object re-id[J]. Advances in Neural Information Processing Systems, 2020, 33: 11309-11321.Google Scholar
Index Terms
- Unsupervised Multi-Source Domain Adaptation for Pedestrian Re-identification ??A Study in Noise-Resilient Learning
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 cross-domain person re-identification by instance and distribution alignment
Highlights- A novel idea of exploring instance-wise localised source knowledge for unsupervised cross-domain person re-id.
AbstractMost existing person re-identification (re-id) methods assume supervised model training on a separate large set of training samples from the target domain. While performing well in the training domain, such trained models are seldom ...
Iterative Discriminative Domain Adaptation
Pattern Recognition and Computer VisionAbstractA popular formulation of domain adaptation (DA) is to simultaneously minimize the source risk and the cross-domain discrepancy between the source domain and target domain . However, this is believed to be suboptimal since the shared feature, ...
Comments