基于Transformer改进的两分支行人重识别算法

doi:10.12068/j.issn.1005-3026.2023.01.004

摘要/Abstract

摘要： 针对基于卷积神经网络的行人重识别算法全局信息建模不足的问题，分析了卷积操作的局限性，提出一种基于Transformer改进的全局-局部两分支行人重识别算法.首先利用相对位置编码改进多头自注意力机制，并将其嵌入到Resnet50骨干网络中.之后在全局分支中对图像进行空间几何划分并利用Transformer的全局感受野增强抽象特征的提取能力；在局部分支中对Layer_3输出进行降维监督，利用多尺度池化获得更丰富的局部特征.实验结果表明，该算法在公开数据集Market-1501和DukeMTMC-reID上的mAP/Rank-1分别达到了93.45%/95.61%和88.79%/90.35%，相对于单纯基于卷积神经网络的算法，本文算法达到更高的精度.

关键词: 行人重识别；Transformer；卷积神经网络；特征提取；多头自注意力机制

Abstract: In order to solve the problem of insufficient global information modeling of person re-recognition algorithm based on convolutional neural network， the limitation of convolution operation is analyzed， and an improved global-local two-branch person re-recognition algorithm based on Transformer is proposed. Firstly， the multi-headed self-attention mechanism which is embedded in the Resnet50 backbone network is optimized by relative position-coding. After that， the processed image is split into two parts on the global branch geometrically， and the ability of extracting the abstract features is enhanced by the Transformer’s global receptive field. On the local branch， the Layer_3 output is under the supervision of dimensionality reduction while the multi-scale pooling obtains richer local features. The experimental result shows that， on the Market-1501 and DukeMTMC-reID datasets， mAP/Rank-1 of the algorithm reaches 93.45%/95.61% and 88.79%/90.35%， respectively. Compared with the algorithm which is only based on convolutional neural network， higher accuracy is achieved.

Key words: person re-identification; Transformer; convolutional neural network; feature extraction; multi-headed self-attention mechanism

中图分类号:

TP181

刘洋，闫冬梅，孟范伟. 基于Transformer改进的两分支行人重识别算法[J]. 东北大学学报（自然科学版）, 2023, 44(1): 26-32.

LIU Yang， YAN Dong-mei， MENG Fan-wei. Improved Two-Branch Person Re-identification Algorithm Based on Transformer[J]. Journal of Northeastern University(Natural Science), 2023, 44(1): 26-32.

参考文献

[1]罗浩，姜伟，范星，等.基于深度学习的行人重识别研究进展［J］.自动化学报，2019，45(11):2032-2049.(Luo Hao，Jiang Wei，Fan Xing，et al.A survey on deep learning based person re-identification［J］.Acta Automatica Sinica，2019，45(11):2032-2049.)
[2]Sun Y，Zheng L，Yang Y，et al.Beyond part models:person retrieval with refined part pooling (and a strong convolutional baseline)［C］//Proceedings of the European Conference on Computer Vision.Munich:Springer，2018:480-496.
[3]Mahdi M K，Emrah B，Muhittin G，et al.Human semantic parsing for person re-identification［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake:IEEE，2018:1062-1071.
[4]Lin T，Dollar P，Girshick R，et al.Feature pyramid networks for object detection［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Hawaii:IEEE，2017:2117-2125.
[5]Zhang Z，Lan C，Zeng W，et al.Relation-aware global attention for person re-identification［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Seattle:IEEE，2020:3186-3195.
[6]Tay C P，Roy S，Yap K H.AANet:attribute attention network for person re-identifications［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE，2019:7134-7143.
[7]Dosovitskiy A，Beyer L，Kolesnikov A，et al.An image is worth 16x16 words:Transformers for image recognition at scale［J］.arXiv preprint arXiv，2020:2010.11929.
[8]Liu Z，Lin Y，Cao Y，et al.Swin transformer:hierarchical vision transformer using shifted windows［J］.arXiv preprint arXiv，2021:2103.14030.
[9]Vaswani A，Shazeer N，Parmar N，et al.Attention is all you need［C］//Advances in Neural Information Processing Systems.La Jolla:NIPS，2017:5998-6008.
[10]Srinivas A，Lin T Y，Parmar N，et al.Bottleneck transformers for visual recognition［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Kuala Lumpur:IEEE，2021:16519-16529.
[11]Szegedy C，Vanhoucke V，Ioffe S，et al.Rethinking the inception architecture for computer vision［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Las Vegas:IEEE，2016:2818-2826.
[12]Lyu Y，Gu Y，Liu X G.The dilemma of TriHard loss and an element-weighted TriHard loss for person re-identification［J］.Advances in Neural Information Processing Systems，2020，33:17391-17402.
[13]Zhong Z，Zheng L，Kang G，et al.Random erasing data augmentation［C］//Proceedings of the AAAI Conference on Artificial Intelligence.Palo Alto :AAAI，2020:13001-13008.
[14]Zheng L，Shen L，Tian L，et al.Scalable person re-identification:a benchmark［C］//Proceedings of the IEEE International Conference on Computer Vision.Santiago:IEEE，2015:1116-1124.
[15]Ristani E，Solera F，Zou R，et al.Performance measures and a data set for multi-target，multi-camera tracking［C］//European Conference on Computer Vision.Amsterdam:Springer，2016:17-35.
[16]Shaw P，Uszkoreit J，Vaswani A.Self-attention with relative position representations［J］.arXiv preprint arXiv，2018:1803.02155.
[17]Wang C，Zhang Q，Huang C，et al.Mancs:a multi-task attentional network with curriculum sampling for person re-identification［C］//Proceedings of the European Conference on Computer Vision.Munich:Springer，2018:365-381.
[18]Hou R，Ma B，Chang H，et al.Interaction-and-aggregation network for person re-identification［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE，2019:9317-9326.
[19]Zheng M，Karanam S，Wu Z，et al.Re-identification with consistent attentive siamese networks［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE，2019:5735-5744.
[20]Zhong Z，Liang Z，Cao D，et al.Re-ranking person re-identification with k-reciprocal encoding［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Hawaii:IEEE，2017:1318-1327.
[21]Si J，Zhang H，Li C G，et al.Dual attention matching network for context-aware feature sequence based person re-identification［C］//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.Salt Lake:IEEE，2018:5363-5372.
[22]Jin X，Lan C，Zeng W，et al.Style normalization and restitution for generalizable person re-identification［C］//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle:IEEE，2020:3143-3152.