结合注意力和纹理特征增强的行人再识别

doi:10.3778/j.issn.1673-9418.2010046

摘要/Abstract

摘要：

针对现有行人再识别算法在处理图像分辨率低、光照差异、姿态和视角多样等情况时,准确率低的问题,提出了基于空间注意力和纹理特征增强的多任务行人再识别算法。算法设计的空间注意力模块更注重与行人属性相关的潜在图像区域,融入属性识别网络,实现属性特征的挖掘;提出的行人再识别网络的纹理特征增强模块通过融合不同空间级别所对应的全局和局部特征,减弱了光照、遮挡等对行人再识别的干扰;最后通过多任务加权损失函数将属性特征和行人特征巧妙融合,避免了由属性异质性造成的再识别精度损失。实验结果表明,该方法在Market-1501和DukeMTMC-reID两大主流行人再识别数据集上的平均精度分别可以达到81.1%和70.1%。

关键词: 空间注意力, 纹理特征增强, 行人属性, 行人再识别

Abstract:

In view of the low accuracy of existing person re-identification to deal with low image resolution, illuminative difference, posture and perspective diversity, this paper proposes a multi-task pedestrian recognition algorithm based on spatial attention and texture feature enhancement. The spatial attention module designed by the algorithm pays more attention to the potential image areas related to the pedestrian attributes, which further explores attribute features. The texture feature enhancement module of the person re-identification network reduces the interference of light, occlusion on person re-identification by fusing the global and local features corresponding to different spatial levels. Finally, the multi-stage weighted loss function integrates the attribute features and pedestrian features to avoid the decrease of mean average precision caused by attribute heterogeneity. Experimental results show that the mean average precision can achieve 81.1% and 70.1% respectively on the Market-1501 and DukeMTMC-reID datasets.

Key words: spatial attention, texture feature enhancement, pedestrian attributes, person re-identification

中图分类号:

TP391

李杰. 结合注意力和纹理特征增强的行人再识别[J]. 计算机科学与探索, 2022, 16(3): 661-668.

LI Jie. Attention and Texture Feature Enhancement for Person Re-identification[J]. Journal of Frontiers of Computer Science and Technology, 2022, 16(3): 661-668.

图/表 7

参考文献 28

[1]	FELZENSZWALB P F, GIRSHICK R B, MCALLESTER D A, et al. Object detection with discriminatively trained part-based models[J]. IEEE Transactions on Software Engineering, 2010, 32(9): 1627-1645.
[2]	GRAY D, TAO H. Viewpoint invariant pedestrian recognition with an ensemble of localized features[C]// LNCS 5302: Proceedings of the 10th European Conference on Computer Vision, Marseille, Oct 12-18, 2008. Berlin, Heidelberg: Springer, 2008: 262-275.
[3]	LOY C C, LIU C X, GONG S G. Person re-identification by manifold ranking[C]// Proceedings of the 2013 IEEE International Conference on Image Processing, Melbourne, Sep 15-18, 2013. Piscataway: IEEE, 2013: 3567-3571.
[4]	谢林江, 季桂树, 彭清, 等. 改进的卷积神经网络在行人检测中的应用[J]. 计算机科学与探索, 2018, 12(5): 708-718.
	XIE L J, JI G S, PENG Q, et al. Application of preprocessing convolutional neural network in pedestrian detection[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(5): 708-718.
[5]	宋丽丽. 迁移度量学习行人再识别算法[J]. 计算机工程与应用, 2019, 55(20): 170-176.
	SONG L L. Transfer metric learning for person re-identification[J]. Computer Engineering and Applications, 2019, 55(20): 170-176.
[6]	宋丽丽, 李彬, 赵俊雅, 等. 正态重采样的改进行人再识别度量学习算法[J]. 计算机工程与应用, 2020, 56(8): 158-165.
	SONG L L, LI B, ZHAO J Y, et al. Normality resampling of improved metric learning method for person re-identification[J]. Computer Engineering and Applications, 2020, 56(8): 158-165.
[7]	LI W, ZHAO R, XIAO T, et al. DeepReID: deep filter pairing neural network for person re-identification[C]// Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, Jun 23-28, 2014. Washington: IEEE Computer Society, 2014: 152-159.
[8]	LIN Y T, ZHENG L, ZHENG Z D, et al. Improving person re-identification by attribute and identity learning[J]. Pattern Recognition, 2019, 95: 151-161. DOI URL
[9]	ZHENG L, SHEN L Y, TIAN L, et al. Scalable person re-identification: a benchmark[C]// Proceedings of the 2015 IEEE International Conference on Computer Vision, Santiago, Dec 7-13, 2015. Washington: IEEE Computer Society, 2015: 1116-1124.
[10]	SCHUMANN A, STIEFELHAGEN R. Person re-identification by deep learning attribute-complementary information[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 1435-1443.
[11]	LIU H, WU J J, JIANG J G, et al. Sequence-based person attribute recognition with joint CTC-Attention model[J]. arXiv:1811.08115, 2018.
[12]	HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]// Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, Jun 27-30, 2016. Washington: IEEE Computer Society, 2016: 770-778.
[13]	HUANG G, LIU Z, VAN DER MAATEN L, et al. Densely connected convolutional networks[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 2261-2269.
[14]	ZHENG Z D, ZHENG L, YANG Y, et al. Unlabeled samples generated by GAN improve the person re-identification baseline in vitro[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 3774-3782.
[15]	FU C Y, LIU W, RANGA A, et al. DSSD: deconvolutional single shot detector[J]. arXiv:1701.06659, 2017.
[16]	ABHINAV S, RAHUL S, JITENDRA M, et al. Beyond skip connections: top-down modulation for object detection[J]. arXiv:1612.06851, 2016.
[17]	DAI J F, QI H Z, XIONG Y W, et al. Deformable convolutional networks[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 764-773.
[18]	FU Y, WEI Y C, ZHOU Y Q, et al. Horizontal pyramid matching for person re-identification[C]// Proceedings of the 33rd AAAI Conference on Artificial Intelligence, the 31st Innovative Applications of Artificial Intelligence Conference, the 9th AAAI Symposium on Educational Advances in Artificial Intelligence, Honolulu, Jan 27-Feb 1, 2019. Menlo Park: AAAI, 2019: 8295-8302.
[19]	LIU J W, ZHA Z J, XIE H T, et al. CA3Net: contextual- attentional attribute-appearance network for person re- identification[C]// Proceedings of the 2018 ACM Multimedia Conference on Multimedia Conference, Seoul, Oct 22-26, 2018. New York: ACM, 2018: 737-745.
[20]	LIN T Y, DOLLÁR P, GIRSHICK R B, et al. Feature pyramid networks for object detection[C]// Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, Jul 21-26, 2017. Washington: IEEE Computer Society, 2017: 936-944.
[21]	SUN Y F, ZHENG L, DENG W J, et al. SVDNet for pedestrian retrieval[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 3820-3828.
[22]	ZHENG Z D, ZHENG J, YANG Y. Pedestrian alignment network for large-scale person re-identification[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2019(10): 3037-3045.
[23]	ZHAO L M, LI X, ZHUANG Y T, et al. Deeply-learned part-aligned representations for person re-identification[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 3219-3228.
[24]	LI W, ZHU X T, GONG S G. Person re-identification by deep joint learning of multi-loss classification[C]// Proceedings of the 26th International Joint Conference on Artificial Intelligence, Melbourne, Aug 19-25, 2017: 2194-2200.
[25]	HERMANS A, BEYER L, LEIBE B. In defense of the triplet loss for person re-identification[J]. arXiv:1703.07737, 2017.
[26]	CHEN Y B, ZHU X T, GONG S G. Person re-identification by deep learning multi-scale representations[C]// Proceedings of the 2017 IEEE International Conference on Computer Vision Workshops, Venice, Oct 22-29, 2017. Washington: IEEE Computer Society, 2017: 2590-2600.
[27]	SUN Y F, ZHENG L, YANG Y, et al. Beyond part models: person retrieval with refined part pooling (and a strong convolutional baseline)[C]// LNCS 11208: Proceedings of the 15th European Conference on Computer Vision, Munich, Sep 8-14, 2018. Cham: Springer, 2018: 501-518.
[28]	LI W, ZHU X T, GONG S G. Harmonious attention network for person re-identification[C]// Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, Jun 18-22, 2018. Washington: IEEE Computer Society, 2018: 2285-2294.

方法	gender	hat	boots	L.up	B.pack	H.bag	Bag	C.shoes	C.up	C.low	mAP
不包含空间注意力模块	69.4	79.4	78.5	87.6	67.7	89.8	83.2	88.6	80.4	90.1	81.5
包含空间注意力模块	86.3	89.9	90.7	88.4	83.4	93.5	81.0	91.2	94.3	91.4	89.0

方法	gender	hat	boots	L.up	B.pack	H.bag	Bag	C.shoes	C.up	C.low	mAP
不包含空间注意力模块	69.4	79.4	78.5	87.6	67.7	89.8	83.2	88.6	80.4	90.1	81.5
包含空间注意力模块	86.3	89.9	90.7	88.4	83.4	93.5	81.0	91.2	94.3	91.4	89.0

λ	Market-1501		DukeMTMC
λ	Rank-1/%	mAP/%	Rank-1/%	mAP/%
2	93.8	80.3	84.4	70.2
4	94.2	81.1	84.7	70.1
6	94.2	80.8	83.6	69.1
8	93.7	80.3	84.1	68.8

λ	Market-1501		DukeMTMC
λ	Rank-1/%	mAP/%	Rank-1/%	mAP/%
2	93.8	80.3	84.4	70.2
4	94.2	81.1	84.7	70.1
6	94.2	80.8	83.6	69.1
8	93.7	80.3	84.1	68.8

实验	方法	Market-1501		DukeMTMC-reID
实验	方法	Rank-1/%	mAP/%	Rank-1/%	mAP/%
1	不包含空间注意力模块的属性识别网络	92.4	79.1	82.6	67.9
2	不包含纹理特征增强模块的行人再识别网络	92.3	78.4	78.0	66.1
3	单任务属性识别网络	74.5	50.2	61.1	38.0
4	单任务行人再识别网络	91.6	77.3	81.1	65.1
5	本文方法	94.2	81.1	84.7	70.1