结合特征融合和通道注意力的多分支换装行人重识别

doi:10.19678/j.issn.1000-3428.0068392

摘要/Abstract

摘要：

换装行人重识别(CC Re-ID)是行人重识别中的一个新兴研究课题, 旨在找出被换衣的行人。当前方法主要集中在使用多模态数据辅助解耦表征学习, 如通过脸、步态、身体轮廓等辅助数据解耦行人自身属性以减少服装影响, 但这些方法泛化能力较差, 需要大量额外工作。此外, 仅使用原始数据的方法对于相关信息的提取不够充分, 性能较弱。针对CC Re-ID存在的上述问题, 提出一种结合特征融合和通道注意力的多分支换装行人重识别方法(MBFC)。通过在主干网络中融入通道注意力机制, 在特征通道层面学习关键信息, 设计局部与全局特征融合方法以提高网络对行人细粒度特征的提取能力。此外, MBFC模型采用多分支结构, 使用服装对抗损失、交叉熵标签平滑损失等多种损失函数引导模型学习与服装无关的信息, 减少服装对模型的影响, 从而提取到更有效的行人信息。在PRCC和VC-Clothes数据集上进行广泛实验, 结果表明, 所提模型在RANK-1和平均精度均值(mAP)指标上优于对比的CC Re-ID方法。

关键词: 换装行人重识别, 多分支, 通道注意力, 特征融合, 注意力机制

Abstract:

Clothes-Changing Person Re-Identification (CC Re-ID) is an emerging research topic in person re-identification, which aims to retrieve pedestrians who have changed their clothes. To date, this task has not been thoroughly studied. Currently, the proposed methods mainly focus on using multi-modal data to assist in decoupling representation learning, such as decoupling the attributes of a pedestrian through auxiliary data such as face, gait, and body contours to reduce the influence of clothing; however, the generalization ability is poor, and additional work is needed to obtain auxiliary information. Furthermore, a method that uses only the original data is insufficient for extracting relevant information, and the performance of the model is poor. To solve the problem of CC Re-ID, a novel multi-branch CC Re-ID method combining feature fusion and channel attention, MBFC, is proposed. This method integrates the channel attention mechanism into the backbone network to learn key information at the feature channel level and designs local and global feature fusion methods to improve the ability of the network to extract fine-grained pedestrian features. In addition, the model adopts a multi-branch structure and uses multiple loss functions, such as clothing counter loss and smooth label cross-entropy loss, to guide the model in learning information unrelated to clothing, reduce the influence of clothing on the model, and thus extract more effective pedestrian information. The proposed model is extensively tested on the PRCC and VC-Clothes datasets. The experimental results indicate that the performance of the proposed model is superior to that of the most advanced CC Re-ID methods in terms of RANK-1 and mean Average Precision (mAP).

Key words: Clothes-Changing Person Re-Identification (CC Re-ID), multi-branch, channel attention, feature fusion, attention mechanism

胡涌涛, 黄洪琼. 结合特征融合和通道注意力的多分支换装行人重识别[J]. 计算机工程, 2025, 51(1): 225-234.

HU Yongtao, HUANG Hongqiong. Multi-Branch Clothes-Changing Person Re-Identification with Feature Fusion and Channel Attention[J]. Computer Engineering, 2025, 51(1): 225-234.

https://www.ecice06.com/CN/Y2025/V51/I1/225

图/表 12

图1 MBFC模型整体结构

Fig.1 Overall structure of MBFC model

图2 特征融合过程

Fig.2 Feature fusion process

图3 SE注意力机制及嵌入方式

Fig.3 SE attention mechanism and embedding mode

图4 图片识别的可视化结果

Fig.4 Visualization results of image recognition

图5 特征融合分支在VC-Clothes数据集上的性能对比

Fig.5 Performance comparison of the feature fusion branch on the VC-Clothes dataset

图6 通道乱序模块在PRCC上的性能对比

Fig.6 Performance comparison of channel out-of-order module on PRCC

参考文献 35

1	ZHENG L, YANG Y, TIAN Q. SIFT meets CNN: a decade survey of instance retrieval. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40(5): 1224- 1244. doi: 10.1109/TPAMI.2017.2709749
2	KARANAM S, GOU M, WU Z, et al. A systematic evaluation and benchmark for person re-identification: features, metrics, and datasets. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2019, 41(3): 523- 536. doi: 10.1109/TPAMI.2018.2807450
3	LENG Q M, YE M, TIAN Q. A survey of open-world person re-identification. IEEE Transactions on Circuits and Systems for Video Technology, 2020, 30(4): 1092- 1108. doi: 10.1109/TCSVT.2019.2898940
4	罗浩, 姜伟, 范星, 等. 基于深度学习的行人重识别研究进展. 自动化学报, 2019, 45(11): 2032- 2049.
	LUO H, JIANG W, FAN X, et al. A survey on deep learning based person re-identification. Acta Automatica Sinica, 2019, 45(11): 2032- 2049.
5	GU X Q, CHANG H, MA B P, et al. Appearance-preserving 3D convolution for video-based person re-identification[EB/OL]. [2023-08-05]. https://arxiv.org/abs/2007.08434.
6	郭业才, 沈宇慧. 融合交互性特征信息的余弦度量行人重识别. 计算机工程与设计, 2023, 44(11): 3395- 3401.
	GUO Y C, SHEN Y H. Person re-identification with cosine metric fusing interactive feature information. Computer Engineering and Design, 2023, 44(11): 3395- 3401.
7	SUN Y F, ZHENG L, YANG Y, et al. Beyond part models: person retrieval with refined part pooling (and a strong convolutional baseline)[EB/OL]. [2023-08-05]. https://arxiv.org/abs/1711.09349.
8	HUANG Y, XU J S, WU Q, et al. Beyond scalar neuron: adopting vector-neuron capsules for long-term person re-identification. IEEE Transactions on Circuits and Systems for Video Technology, 2020, 30(10): 3459- 3471. doi: 10.1109/TCSVT.2019.2948093
9	SHU X J, WANG X, ZANG X H, et al. Large-scale spatio-temporal person re-identification: algorithms and benchmark. IEEE Transactions on Circuits and Systems for Video Technology, 2022, 32(7): 4390- 4403. doi: 10.1109/TCSVT.2021.3128214
10	张鹏, 张晓林, 包永堂, 等. 换装行人重识别研究进展. 中国图象图形学报, 2023, 28(5): 1242- 1264.
	ZHANG P, ZHANG X L, BAO Y T, et al. Cloth-changing person re-identification: a summary. Journal of Image and Graphics, 2023, 28(5): 1242- 1264.
11	CHANG X B, HOSPEDALES T M, XIANG T. Multi-level factorisation net for person re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2018: 2109-2118.
12	WAN F B, WU Y, QIAN X L, et al. When person re-identification meets changing clothes[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 3620-3628.
13	YU S J, LI S H, CHEN D P, et al. COCAS: a large-scale clothes changing person dataset for re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 3400-3409.
14	FAN L J, LI T H, FANG R Y, et al. Learning longterm representations for person re-identification using radio signals[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 1-2.
15	WANG Y X, DU B W, SHEN Y R, et al. EV-gait: event-based robust gait recognition using dynamic vision sensors[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2019: 6358-6367.
16	FAN C, PENG Y J, CAO C S, et al. GaitPart: temporal part-based model for gait recognition[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 14213-14221.
17	JIN X, HE T Y, ZHENG K C, et al. Cloth-changing person re-identification from a single image with gait prediction and regularization[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2022: 14258-14267.
18	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2016: 770-778.
19	GU X Q, CHANG H, MA B P, et al. Clothes-changing person re-identification with RGB modality only[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2022: 1050-1059.
20	SHU X J, LI G, WANG X, et al. Semantic-guided pixel sampling for cloth-changing person re-identification. IEEE Signal Processing Letters, 2021, 28, 1365- 1369. doi: 10.1109/LSP.2021.3091924
21	HUANG Y, WU Q, XU J S, et al. Clothing status awareness for long-term person re-identification[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. Washington D.C., USA: IEEE Press, 2021: 11875-11884.
22	HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011- 2023. doi: 10.1109/TPAMI.2019.2913372
23	YANG Q, WU A, ZHENG W S. Person re-identification by contour sketch under moderate clothing change. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2021, 43(6): 2029- 2046. doi: 10.1109/TPAMI.2019.2960509
24	DENG J, DONG W, SOCHER R, et al. ImageNet: a large-scale hierarchical image database[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2009: 248-255.
25	LI W, ZHU X T, GONG S G. Harmonious attention network for person re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2018: 2285-2294.
26	QIAN X L, FU Y W, JIANG Y G, et al. Multi-scale deep learning architectures for person re-identification[C]//Proceedings of the IEEE International Conference on Computer Vision. Washington D.C., USA: IEEE Press, 2017: 5409-5418.
27	KOSTINGER M, HIRZER M, WOHLHART P, et al. Large scale metric learning from equivalence constraints[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2012: 2288-2295.
28	LIAO S C, HU Y, ZHU X Y, et al. Person re-identification by local maximal occurrence representation and metric learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2015: 2197-2206.
29	CHEN J X, JIANG X Y, WANG F D, et al. Learning 3D shape feature for texture-insensitive person re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2021: 8142-8151.
30	HONG P X, WU T, WU A C, et al. Fine-grained shape-appearance mutual learning for cloth-changing person re-identification[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2021: 10508-10517.
31	ZHANG Q L, YANG Y B. SA-Net: Shuffle attention for deep convolutional neural networks[C]//Proceedings of the ICASSP IEEE International Conference on Acoustics, Speech and Signal Processing. Washington D.C., USA: IEEE Press, 2021: 2235-2239.
32	LIU Y C, SHAO Z R, HOFFMANN N. Global attention mechanism: retain information to enhance channel-spatial interactions[EB/OL]. [2023-08-05]. https://arxiv.org/abs/2112.05561.
33	WOO S, PARK J, LEE J Y, et al. CBAM: convolutional block attention module[EB/OL]. [2023-08-05]. https://link.springer.com/chapter/10.1007/978-3-030-01234-2_1.
34	MISRA D, NALAMADA T, ARASANIPALAI A U, et al. Rotate to attend: convolutional triplet attention module[C]//Proceedings of the IEEE Winter Conference on Applications of Computer Vision. Washington D.C., USA: IEEE Press, 2021: 3138-3147.
35	WANG Q L, WU B G, ZHU P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington D.C., USA: IEEE Press, 2020: 11534-11542.

[1]	罗旭东, 袁笛, 常晓军, 何震宇. 基于不确定性启发图像增强的水下目标跟踪[J]. 计算机工程, 2025, 51(1): 11-19.
[2]	周宇, 谢威, 邝得互, 江健民. 基于三元自注意力的视频快照压缩成像重建[J]. 计算机工程, 2025, 51(1): 20-30.
[3]	费涛, 艾山·吾买尔, 杜文旭, 朱翠翠. 基于Squeezeformer的多颗粒度多方面发音质量评测方法[J]. 计算机工程, 2025, 51(1): 81-87.
[4]	周雪阳, 傅启明, 陈建平, 陈延明, 陆悠, 王蕴哲. 基于证据和图推理的文档级关系抽取方法: 以医学关系为例[J]. 计算机工程, 2025, 51(1): 106-117.
[5]	王翔, 魏玉锌, 毛国君. 一种融合图数据多元结构和特征的图池化方法[J]. 计算机工程, 2025, 51(1): 128-137.
[6]	肖超恩, 李子凡, 张磊, 王建新, 钱思源. 基于Transformer模型与注意力机制的差分密码分析[J]. 计算机工程, 2025, 51(1): 156-163.
[7]	杨红菊, 吉昌. 学习驱动的图像压缩算法研究[J]. 计算机工程, 2025, 51(1): 190-197.
[8]	火久元, 苏泓瑞, 武泽宇, 王婷娟. 基于改进YOLOv8的道路交通小目标车辆检测算法[J]. 计算机工程, 2025, 51(1): 246-257.
[9]	郑雅洲, 刘万平, 黄东. 一种基于注意力机制的BERT-CNN-GRU检测方法[J]. 计算机工程, 2025, 51(1): 258-268.
[10]	王骞, 张俊华, 王泽彤, 李博. X2S-Net:基于双平面X线片的脊柱三维重建[J]. 计算机工程, 2025, 51(1): 277-286.
[11]	李猛坤, 袁晨, 王琪, 赵冲, 陈景轩, 刘立峰. 基于改进YOLOv8算法的在线听课行为识别模型研究[J]. 计算机工程, 2025, 51(1): 287-294.
[12]	刘钟, 唐宏, 王宁喆, 朱传润. 融合RNN与稀疏自注意力的文本摘要方法[J]. 计算机工程, 2025, 51(1): 312-320.
[13]	刘兆伟, 方艳红, 郑明宇, 锁斌. 基于注意力机制与多任务的肺部疾病诊断方法[J]. 计算机工程, 2025, 51(1): 332-342.
[14]	李俊俊, 董建刚, 李坤. 基于Kubernetes的集群节能策略研究[J]. 计算机工程, 2024, 50(9): 82-91.
[15]	林畅, 郭伟, 任哲聪, 金海波. 基于Transformer的目标跟踪与分割统一算法[J]. 计算机工程, 2024, 50(9): 130-141.

选择文件类型/文献管理软件名称

选择包含的内容