基于强区分性特征挖掘的无监督行人重识别
Unsupervised Person Re-Identification Based on Discriminative Feature Mining
DOI: 10.12677/mos.2024.133186, PDF,   
作者: 徐 锟, 王晓红, 赵 徐, 徐世豪：上海理工大学出版印刷与艺术设计学院，上海 收稿日期：2024年3月28日；录用日期：2024年5月2日；发布日期：2024年5月8日
关键词: 特征挖掘；无监督学习；伪标签；聚类；Feature Mining； Unsupervised Learning； Pseudo Labels； Clusters
摘要: 目前，基于深度学习的无监督行人重识别正在通过改进聚类方法来提高生成伪标签的准确性和鲁棒性。然而，行人的固有物理特征，如肢体和身体，尚未充分利用。此外，遮挡和姿势导致了在行人图像中准确匹配和识别局部区域的困难，而行人服装颜色引起的噪声干扰也对行人重识别准确性产生了重要影响。为了解决这些问题，本文提出了一种判别性特征挖掘网络(DFMN)，通过引入Transformer注意力机制来突出显示行人的有效物理特征；为了深入挖掘不同样本之间的细粒度局部特征，本文采用了基于最短路径的部分对齐分割机制；同时，针对行人服装颜色引起的冗余信息，本文融合了一个alpha通道，可以有效消除噪声干扰。实验结果表明，本文所提出的方法在Market1501数据集和MSMT17数据集上的map指标上分别实现了4.0%和6.4%的提升。
Abstract: At present, Unsupervised Person Re-Identification based on deep learning promotes the accuracy and robustness of generating pseudo labels mainly by improving the clustering method. However, for inherent physical features of pedestrians, such as limbs, and bodies, which are not fully utilized, occlusion, posture pose difficulties in accurately matching and recognizing local areas in pedestrian images ,and noise interference caused by pedestrian clothing color have essential impact on person re-identification accuracy. In this paper, we propose a Discriminative Feature Minning Network (DFMN) by introducing a Transformer attention mechanism to highlight the pedestrians’ effective physical features. To deeply mining the fine-grained local features among different samples, a part-aligned segmentation mechanism based on the shortest path is used. At the same time, as for the redundant information resulting from the color of pedestrian clothing, we incorporate an alpha channel, which can effectively eliminate noise interference. The experimental results show that the proposed method has achieved improvements of 4.0% and 6.4% on the Market1501 dataset and the MSMT17 dataset for the map metric.
文章引用：徐锟, 王晓红, 赵徐, 徐世豪. 基于强区分性特征挖掘的无监督行人重识别[J]. 建模与仿真, 2024, 13(3): 2011-2022. https://doi.org/10.12677/mos.2024.133186

参考文献

[1] Deng, W.J., Zheng, L., Ye, Q.X., et al. (2018) Image-Image Domain Adaptation with Preserved Self-Similarity and Domain-Dissimilarity for Person Re-identification. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 18-23 June 2018, 994-1003. [Google Scholar] [CrossRef] 
[2] Zhong, Z., Zheng, L., Li, S.Z., et al. (2018) Generalizing a Person Retrieval Model Hetero-and Homogeneously. Computer Vision-ECCV 2018, Munich, 8-14 September 2018, 176-192. [Google Scholar] [CrossRef] 
[3] Zhong, Z., Zheng, L., Luo, Z.M., et al. (2019) Invariance Matters: Exemplar Memory for Domain Adaptive Person Re-identification. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 15-20 June 2019, 598-607. [Google Scholar] [CrossRef] 
[4] Wei, L., Zhang, S., Gao, W. and Tian, Q. (2018) Person Transfer GAN to Bridge Domain Gap for Person Re-Identification. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 18-23 June 2018, 79-88. [Google Scholar] [CrossRef] 
[5] Wu, J., Yang, Y., Liu, H., Liao, S., Lei, Z. and Li, S.Z. (2019) Unsupervised Graph Association for Person Re-Identification. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, 27 October-2 November 2019, 8321-8330. [Google Scholar] [CrossRef] 
[6] Wu, Z., Xiong, Y., Yu, S.X. and Lin, D. (2018) Unsupervised Feature Learning via Non-Parametric Instance Discrimination. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 18-23 June 2018, 3733-3742. [Google Scholar] [CrossRef] 
[7] Fan, H., Zheng, L., Yan, C. and Yang, Y. (2018) Unsupervised Person Re-Identification: Clustering and Fine-Tuning. ACM Transactions on Multimedia Computing, Communications, and Applications, 14, Article No. 83. [Google Scholar] [CrossRef] 
[8] Fu, Y., Wei, Y., Wang, G., Zhou, Y., Shi, H. and Huang, T.S. (2019) Self-Similarity Grouping: A Simple Unsupervised Cross Domain Adaptation Approach for Person Re-Identification. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), Seoul, 27 October-2 November 2019, 6112-6121. [Google Scholar] [CrossRef] 
[9] Lin, Y., Dong, X., Zheng, L., Yan, Y. and Yang, Y. (2019) A Bottom-Up Clustering Approach to Unsupervised Person Re-identification. Proceedings of the AAAI Conference on Artificial Intelligence, 33, 8738-8745. [Google Scholar] [CrossRef] 
[10] Wang, D. and Zhang, S. (2020) Unsupervised Person Re-Identification via Multi-Label Classification. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, 13-19 June 2020, 10981-10990. [Google Scholar] [CrossRef] 
[11] Ge, Y., Zhu, F., Chen, D., Zhao, R. and Li, H.S. (2020) Self-Paced Contrastive Learning with Hybrid Memory for Domain Adaptive Object Re-ID. Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver, 6-12 December 2020, 11309-11321.
[12] Ester, M., Kriegel, H.-P., Sander, J., Xu, X., et al. (1996) A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise. Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, Portland Oregon, 2-4 August 1996, 226-231.
[13] MacQueen, J., et al. (1967) Some Methods for Classification and Analysis of Multivariate Observations. Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Oakland, 27 December 1965-7 January 1966, 281-297.
[14] Rosvall, M. and Bergstrom, C.T. (2008) Maps of Random Walks on Complex Networks Reveal Community Structure. Proceedings of the National Academy of Sciences, 105, 1118-1123. [Google Scholar] [CrossRef] [PubMed]
[15] Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention Is All You Need. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, 4-9 December 2017, 6000-6010.
[16] Han, K., Wang, Y., Chen, H., Chen, X., Guo, J., Liu, Z., Tang, Y., Xiao, A., Xu, C., Xu, Y. et al. (2023) A Survey on Visual Transformer. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45, 87-110. [Google Scholar] [CrossRef] 
[17] Khan, S., Naseer, M., Hayat, M., et al. (2022) Transformers in Vision: A Survey. ACM Computing Surveys, 54, Article No. 200. [Google Scholar] [CrossRef] 
[18] Zheng, L., Huang, Y.J., Lu, H.C., et al. (2019) Pose-Invariant Embedding for Deep Person Re-Identification. IEEE Transactions on Image Processing, 28, 4500-4509. [Google Scholar] [CrossRef] 
[19] Chen, H., Lagadec, B. and Bremond, F. (2021) ICE: Inter-Instance Contrastive Encoding for Unsupervised Person Re-Identification. 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, 10-17 October 2021, 14940-14949. [Google Scholar] [CrossRef] 
[20] Isobe, T., Li, D., Tian, L., Chen, W., Shan Y. and Wang, S. (2021) Towards Discriminative Representation Learning for Unsupervised Person Re-Identification. 2021 IEEE/CVF International Conference on Computer Vision (ICCV), Montreal, 10-17 October 2021, 8506-8516. [Google Scholar] [CrossRef] 
[21] Cho, Y., Kim, W.J., Hong, S., et al. (2022) Part-Based Pseudo Label Refinement for Unsupervised Person Re-Identification. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, 18-24 June 2022, 7298-7308. [Google Scholar] [CrossRef] 
[22] Yu, H.X., Zheng, W.S., Wu, A., Guo, X., Gong, S. and Lai, J.H. (2019) Unsupervised Person Re-Identification by Soft Multilabel Learning. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 15-20 June 2019, 2148-2157. [Google Scholar] [CrossRef] 
[23] Kheradpisheh, S.R., Ganjtabesh, M., Thorpe, S.J., et al., (2018) STDP-Based Spiking Deep Convolutional Neural Networks for Object Recognition. Neural Networks, 99, 56-67. [Google Scholar] [CrossRef] [PubMed]
[24] Wang, P.C., Wang, X., Luo, H., et al. (2022) Scaled ReLU Matters for Training Vision Transformers. Proceedings of the AAAI Conference on Artificial Intelligence, 36, 2495-2503. [Google Scholar] [CrossRef] 
[25] Xiao, T., Singh, M., Mintun, E., et al. (2021) Early Convolutions Help Transformers See Better. 35th Conference on Neural Information Processing Systems (NeurIPS 2021), Beijing, 4 November 2021.
[26] Tong, H.Y. and Ng, K.W. (2021) A Bottom-Up Clustering Approach to Identify Bus Driving Patterns and to Develop Bus Driving Cycles for Hong Kong. Environmental Science and Pollution Research International, 28, 14343-14357. [Google Scholar] [CrossRef] [PubMed]
[27] Lin, Y., Wu, Y., Yan, C., et al. (2020) Unsupervised Person Re-Identification via Cross-Camera Similarity Exploration. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, 13-19 June 2020, 3387-3396. [Google Scholar] [CrossRef] 
[28] Zeng, K., Ning, M., Wang, Y. and Guo, Y. (2020) Hierarchical Clustering with Hard-Batch Triplet Loss for Person Re-Identification. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, 13-19 June 2020, 13654-13662. [Google Scholar] [CrossRef] 
[29] Han, X.M., Yu, X.H., Li, G.R., et al. (2023) Rethinking Sampling Strategies for Unsupervised Person Re-Identification, IEEE Transactions on Image Processing, 32, 29-42. [Google Scholar] [CrossRef] 
[30] Dai, Z., Wang, G., Yuan, W., Zhu, S. and Tan, P. (2022) Cluster Contrast for Unsupervised Person Re-Identification. Computer Vision-ACCV 2022, Macao, 4-8 December 2022, 319-337. [Google Scholar] [CrossRef] 
[31] Li, M., Zhu, X. and Gong, S. (2018) Unsupervised Person Re-Identification by Deep Learning Tracklet Association. Computer Vision-ECCV 2018, Munich, 8-14 September 2018, 772-788. [Google Scholar] [CrossRef] 
[32] Li, M., Zhu, X. and Gong, S. (2019) Unsupervised Tracklet Person Re-Identification. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42, 1770-1782. [Google Scholar] [CrossRef] 

为你推荐