结合MASP和语义分割的双链路行人重识别方法

doi:10.3778/j.issn.1002-8331.2105-0466

摘要/Abstract

摘要： 行人重识别是通过不同的摄像机识别同一个人。由于人的姿势多变，背景杂乱以及拍摄角度不同等，提取强大的行人特征成为一个有挑战性的任务。为了提取良好的行人特征表示，提出了一种结合MASP与语义分割的双链路行人重识别模型。该方法对网络不同深度的特征进行采样，不同深度的特征图具有不同的表达能力，使网络可以学习到行人身上更加细粒度的特征。上层链路针对网络过深导致行人信息丢失的问题，提出了MASP模块，对浅层特征进行采样，然后与高级特征连接，对深浅层级特征交融，增加特征的多样性。下层链路基于语义分割结果，对提取的中间层行人特征映射，得出语义部位特征。在测试阶段，将全局特征与语义部位特征结合生成多层次特征，加强模型的表征能力。在Market-1501和DukeMTMC-reID两个数据集上与其他方法的对比以及消融实验表明，提出的结合MASP与语义分割的双链路重识别模型有效提升行人重识别性能。

关键词: 行人重识别, 双链路, 全局特征, 语义分割, 多空洞空间卷积金字塔（MASP）

Abstract: Person re-identification is the identification of the same person through different cameras. Extracting strong person feature is a challenging task due to the varied postures, messy backgrounds, and different camera angles. In order to extract good pedestrian feature representation, a dual link person re-identification model combining MASP and semantic segmentation is proposed. In view of the problem of person information loss caused by too deep network in the upper link, the MASP module is proposed to blend features of depth and depth levels to increase the diversity of features. Based on the results of semantic segmentation, the lower link maps the features of different parts of persons to obtain the semantic features of different parts. In the test stage, the global features and semantic features are combined to generate multi-level features to enhance the representational ability of the model. The comparison with other methods and ablation experiments on Market-1501 and DukeMTMC-reID data sets show that the proposed dual link re-recognition model combining MASP and semantic segmentation can effectively improve person re-identification performance.

Key words: person re-identification, dual-link, global feature, semantic segmentation, multiple atrous spatial pyramid（MASP）

朱亚梅, 施一萍, 江悦莹, 邓源, 刘瑾. 结合MASP和语义分割的双链路行人重识别方法[J]. 计算机工程与应用, 2022, 58(24): 143-150.

ZHU Yamei, SHI Yiping, JIANG Yueying, DENG Yuan, LIU Jin. Dual-Link Person Re-Identification Method Based on MASP and Semantic Segmentation[J]. Computer Engineering and Applications, 2022, 58(24): 143-150.

参考文献

[1] LI Z，TAO X，GONG S.Learning a discriminative null space for person re-identification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，Las Vegas，NV，USA，Jun 27-30，2016.Piscataway：IEEE，2016：1239-1248.
[2] LIAO S，YANG H，ZHU X，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，Boston，MA，USA，June 12，2015.Piscataway：IEEE，2015：2197-2206.
[3] 郑远攀，李广阳，李晔.深度学习在图像识别中的应用研究综述[J].计算机工程与应用，2019，55（12）：20-36.
ZHENG Y P，LI G Y，LI Y.Survey of application of deep learning in image recognition[J].Computer Engineering and Applications，2019，55（12）：20-36.
[4] 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（ECCV），2018：480-496.
[5] LI D，CHEN X，ZHANG Z，et al.Learning deep context-aware features over body and latent parts for person re-identification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，Honolulu，HI，USA，July 21-26 2017.Piscataway：IEEE，2017：384-393.
[6] CHENG D，GONG Y，ZHOU S，et al.Person re-identification by multi-channel parts-based CNN with improved triplet loss function[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，Las Vegas，NV，USA，Jun 27-30，2016.Piscataway：IEEE，2016：1335-1344.
[7] XIAO Q，LUO H，ZHANG C.Margin sample mining loss：a deep learning based method for person re-identification[J].arXiv：1710.00478，2017.
[8] 肖雅妮，范馨月，陈文峰.多分支融合局部特征的行人重识别算法[J].计算机工程与应用，2021，57（18）：213-219.
XIAO Y N，FAN X Y，CHEN W F.Research on person re-identification based on integrating local features under multi-branches[J].Computer Engineering and Applications，2021，57（18）：213-219.
[9] YU F，KOLTUN V.Multi-scale context aggregation by dilated convolutions[J].arXiv：1511.07122，2015.
[10] LI P，XU Y，WEI Y，et al.Self-correction for human parsing[J].arXiv：1910.09777，2019.
[11] LUO H，JIANG W，GU Y，et al.A strong baseline and batch normalization neck for deep person re-identification[J].IEEE Transactions on Multimedia，2019，22（10）：2597-2609.
[12] 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，NV，USA，June 30，2016.Piscataway：IEEE，2016：2818-2826.
[13] ZHENG L，SHEN L，TIAN L.Scalable person re-identification：a benchmark[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，Santiago，Chile，Dec 7-13，2015.Piscataway：IEEE，2015：1116-1124.
[14] ZHENG Z，ZHENG L，YANG Y.Unlabeled samples generated by GAN improve the person re-identification baseline in vitro[C]//Proceedings of the IEEE International Conference on Computer Vision，Venice，Italy，Oct 22-29，2017.Piscataway：IEEE，2017：3774-3782.
[15] LIANG X，GONG K，SHEN X，et al.Look into person：joint body parsing & pose estimation network and a new benchmark[J].arXiv：1804.01984，2018.
[16] KALAYEH M M，BASARAN E，GOKMEN M，et al.Human semantic parsing for person re-identification[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition，Salt Lake City，UT，USA，June 18-23，2018.Piscataway：IEEE，2018：1062-1071.
[17] SUN Y F，XU Q，LI Y L.Perceive where to focus：learning visibility-aware part-level features for partial person re-identification[C]//Proceedings of 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition，Long Beach，CA，USA，June 15-20，2019.Piscataway：IEEE，2019：393-402.
[18] FU Y，WEI Y C，ZHOU Y Q，et al.Horizontal pyramid matching for person re-identification[C]//Proceedings of the AAAI Conference on Artificial Intelligence，Honolulu，HI，USA，February 1，2019.Palo Alto，CA：AAAI Press，2019：8295-8302.
[19] 刘紫燕，万培佩.基于注意力机制的行人重识别特征提取方法[J].计算机应用，2020，40（3）：672-676.
LIU Z Y，WAN P P.Pedestrian re-identification feature extraction method based on attention mechanism[J].Journal of Computer Applications，2020，40（3）：672-676.
[20] TAY C P，ROY S，YAP K H.AANet：attribute attention network for person re-identifications[C]//Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition，Long Beach，California，USA，June 15-20，2019.Piscataway：IEEE，2019：7127-7136.