融合属性特征的行人重识别方法

邵晓雯; 帅惠; 刘青山

doi:10.16383/j.aas.c190763

融合属性特征的行人重识别方法

doi: 10.16383/j.aas.c190763

1.
南京信息工程大学自动化学院江苏省大数据分析技术重点实验室南京 210044

基金项目: 国家自然科学基金(61532009, 61825601)资助

详细信息

作者简介:
邵晓雯：南京信息工程大学自动化学院硕士研究生. 2018年获得南京信息工程大学电子与信息工程学院学士学位. 主要研究方向为计算机视觉, 行人重识别.E-mail: xiaowen_shao@nuist.edu.cn

帅惠：南京信息工程大学博士研究生. 2018年获得南京信息工程大学信息与控制学院硕士学位. 主要研究方向为目标检测, 3D 场景解析.E-mail: huishuai13@163.com

刘青山：南京信息工程大学自动化学院院长, 教授. 2003年获得中国科学院自动化研究所博士学位. 主要研究方向为图像理解, 模式识别, 机器学习. 本文通信作者.E-mail: qsliu@nuist.edu.cn

计量
- 文章访问数: 907
- HTML全文浏览量: 541
- PDF下载量: 262
- 被引次数: 0
出版历程
- 收稿日期: 2019-11-04
- 录用日期: 2020-04-06
- 网络出版日期: 2021-12-24
- 刊出日期: 2022-02-18

Person Re-identification Based on Fused Attribute Features

1.
Jiangsu Key Laboratory of Big Data Analysis Technology, School of Automation, Nanjing University of Information Science and Technology, Nanjing 210044

Funds: Supported by National Natural Science Foundation of China (61532009, 61825601)

More Information

Author Bio:
SHAO Xiao-Wen　Master student at the School of Automation, Nanjing University of Information Science and Technology. She received her bachelor degree from the School of Electronic and Information Engineering, Nanjing University of Information Science and Technology in 2018. Her research interest covers computer vision and person re-identification

SHUAI Hui　Ph.D. candidate at Nanjing University of Information Science and Technology. He received his master degree from the School of Information and Control, Nanjing University of Information Science and Technology in 2018. His research interest covers object detection and 3D scene analysis

LIU Qing-Shan　Dean and professor of the School of Automation, Nanjing University of Information Science and Technology. He received his Ph.D. degree from the Institute of Automation, Chinese Academy of Sciences in 2003. His research interest covers image understanding, pattern recognition, and machine learning. Corresponding author of this paper

摘要

摘要: 行人重识别旨在跨监控设备下检索出特定的行人目标. 由于不同的行人可能具有相似的外观, 因此要求行人重识别模型能够捕捉到充足的细粒度特征. 本文提出一种融合属性特征的行人重识别的深度网络方法, 将行人重识别和属性识别集成在分类网络中, 进行端到端的多任务学习. 此外, 对于每张输入图片, 网络自适应地生成对应于每个属性的权重, 并将所有属性的特征以加权求和的方式结合起来, 与全局特征一起用于行人重识别任务. 全局特征关注行人的整体外观, 而属性特征关注细节区域, 两者相互补充可以对行人进行更全面的描述. 在行人重识别的主流数据集DukeMTMC-reID和Market-1501上的实验结果表明了本文方法的有效性, 平均精度均值(Mean average precision, mAP)分别达到了74.2%和83.5%, Rank-1值分别达到了87.1%和93.6%. 此外, 在这两个数据集上的属性识别也得到了比较好的结果.
- 行人重识别 /
- 属性识别 /
- 深度学习 /
- 自适应权重
Abstract: Person re-identification aims to retrieve specific pedestrian target across surveillance devices. Since different pedestrians may have a similar appearance, it requires the person re-identification model to capture sufficient fine-grained features. This paper proposes a new deep network method for person re-identification based on fused attribute features. Person re-identification and attribute recognition are integrated into the classification network for end-to-end multi-task learning. In addition, for each input image, the network adaptively generates weights corresponding to each attribute. Features of all attributes are combined in a weighted summation way, and together with the global feature for person re-identification task. Global feature focuses on the overall appearance of the pedestrian, while the attribute feature focuses on the detail area, and they can complement each other to provide a more comprehensive description of the pedestrian. Experimental results on the DukeMTMC-reID and Market-1501 datasets, two popular datasets of person re-identification, show the effectiveness of the proposed method. The mean average precision (mAP) values reach 74.2% and 83.5% respectively, and the Rank-1 values reach 87.1% and 93.6% respectively. In addition, attribute recognition on these two datasets also achieved better results.
- Person re-identification /
- attribute recognition /
- deep learning /
- adaptive weight

HTML全文

图 1 网络结构示意图

Fig. 1 Schematic diagram of network structure

下载: 全尺寸图片幻灯片

图 2 设置不同的$\alpha$和$\beta$的结果

Fig. 2 Results setting different $\alpha$ and $\beta$

下载: 全尺寸图片幻灯片

图 3 各个属性的可视化结果及对应的权重值

Fig. 3 Visualization result and corresponding weight value of each attribute

下载: 全尺寸图片幻灯片

图 4 使用不同特征检索到的图片

Fig. 4 Images retrieved by different features

下载: 全尺寸图片幻灯片

表 1 与相关方法的性能比较(%)

Table 1 Performance comparison with related methods (%)

方法	DukeMTMC-reID		Market-1501
方法	mAP	Rank-1	mAP	Rank-1
PCB-RPP^[13]	69.2	83.3	81.6	93.8
PDC^[14]	—	—	63.4	84.4
PSE^[15]	62.0	79.8	69.0	87.7
SPReID^[16]	71.0	84.4	81.3	92.5
ACRN^[21]	52.0	72.6	62.6	83.6
APR^[22]	55.6	73.9	66.9	87.0
AANet-50^[25]	72.6	86.4	82.5	93.9
本文	74.2	87.1	83.5	93.6

下载: 导出CSV

表 2 使用不同损失函数的性能比较(%)

Table 2 Performance comparison using different loss functions (%)

损失函数	mAP	Rank-1
$L_{{\rm{id}}}$	70.5	84.1
$L_{{\rm{id}}}+\beta L_{{\rm{att}}}$	71.1	85.8
$L_{{\rm{id}}}+\alpha L_{{\rm{local}}}+\beta L_{{\rm{att}}}$	74.2	87.1
${L_{ {\rm{id} } }+\alpha L_{ {\rm{local} } }+\beta L^{*}_{ {\rm{att} } } }$	73.0	86.0
$L_{ {\rm{id} } }\,({\rm{no} } \ {\rm{LS} })$	66.8	83.3

下载: 导出CSV

表 3 使用不同特征融合方式的性能比较(%)

Table 3 Performance comparison using different feature fusion methods (%)

方法	特征维度	mAP	Rank-1
var1	1024	71.6	84.6
var2	1024	72.5	85.7
var3	1024	74.1	86.3
本文	1024	74.2	87.1
特征${\boldsymbol g}$	512	72.4	86.1
特征${\boldsymbol h}_{w}$	512	71.8	85.1

下载: 导出CSV

表 4 DukeMTMC-reID上属性识别的准确率(%)

Table 4 Accuracy of attribute recognition on DukeMTMC-reID (%)

方法	gender	hat	boots	l.up	b.pack	h.bag	bag	c.shoes	c.up	c.low	平均值
APR^[22]	84.2	87.6	87.5	88.4	75.8	93.4	82.9	89.7	74.2	69.9	83.4
B2	85.94	89.75	89.92	88.64	84.35	93.61	83.06	90.70	74.62	66.81	84.74
本文	86.07	90.74	89.72	88.78	84.47	93.28	82.04	91.79	75.53	68.14	85.06

下载: 导出CSV

表 5 Market-1501上属性识别的准确率(%)

Table 5 Accuracy of attribute recognition on Market-1501 (%)

方法	gender	age	hair	l.slv	l.low	s.clth	b.pack	h.bag	bag	hat	c.up	c.low	平均值
APR^[22]	88.9	88.6	84.4	93.6	93.7	92.8	84.9	90.4	76.4	97.1	74.0	73.8	86.6
AANet-50^[25]	92.31	88.21	86.58	94.45	94.24	94.83	87.77	89.61	79.72	98.01	77.08	70.81	87.80
B2	93.05	85.82	88.97	93.56	94.18	94.53	88.80	88.08	79.24	98.28	74.95	68.64	87.34
本文	93.69	86.16	89.00	94.11	94.80	94.81	89.33	88.68	79.03	98.33	76.52	70.68	87.93

下载: 导出CSV

参考文献(31)

[1]	Dalal N, Triggs B. Histograms of oriented gradients for human detection. In: Proceedings of the 2005 IEEE Conference on Computer Vision and Pattern Recognition. San Diego, CA, USA: IEEE, 2005. 886−893
[2]	Weinberger K Q, Saul L K. Fast solvers and efficient implementations for distance metric learning. In: Proceedings of the 25th International Conference on Machine Learning. Helsinki, Finland: ACM, 2008. 1160−1167
[3]	Liao S C, Hu Y, Zhu X Y, Li S Z. Person re-identification by local maximal occurrence representation and metric learning. In: Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Boston, MA, USA: IEEE, 2015. 2197−2206
[4]	Ma F, Jing X Y, Zhu X K, Tang Z M, Peng Z P. True-color and grayscale video person re-identification. IEEE Transactions on Information Forensics and Security, 2020, 15: 115−129 doi: 10.1109/TIFS.2019.2917160
[5]	Ma F, Zhu X K, Zhang X Y, Yang L, Zuo M, Jing X Y. Low illumination person re-identification. Multimedia Tools and Applications, 2019, 78(1): 337−362 doi: 10.1007/s11042-018-6239-3
[6]	Ma F, Zhu X K, Liu Q L, Song C F, Jing X Y, Ye D P. Multi-view coupled dictionary learning for person re-identification. Neurocomputing, 2019, 348: 16−26 doi: 10.1016/j.neucom.2018.07.081
[7]	(罗浩, 姜伟, 范星, 张思朋. 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 Hao, Jiang Wei, Fan Xing, Zhang Si-Peng. A survey on deep learning based person re-identification. Acta Automatica Sinica, 2019, 45(11): 2032−2049
[8]	Jing X Y, Zhu X K, Wu F, Hu R M, You X G, Wang Y H, et al. Super-resolution person re-identification with semi-coupled low-rank discriminant dictionary learning. IEEE Transactions on Image Processing, 2017, 26(3): 1363−1378 doi: 10.1109/TIP.2017.2651364
[9]	Hermans A, Beyer L, Leibe B. In defense of the triplet loss for person re-identification. arXiv preprint, arXiv: 1703.07737, 2017.
[10]	Chen W H, Chen X T, Zhang J G, Huang K Q. Beyond triplet loss: A deep quadruplet network for person reidentification. In: Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, Hawaii, USA: IEEE, 2017. 403−412
[11]	Chen D P, Xu D, Li H S, Sebe N, Wang X G. Group consistent similarity learning via deep crf for person reidentification. In: Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, Utah, USA: IEEE, 2018. 8649−8658
[12]	Zhu X K, Jing X Y, Zhang F, Zhang X Y, You X G, Cui X. Distance learning by mining hard and easy negative samples for person re-identification. Pattern Recognition, 2019, 95: 211−222 doi: 10.1016/j.patcog.2019.06.007
[13]	Sun Y F, Zheng L, Yang Y, Tian Q, Wang S J. Beyond part models: Person retrieval with refined part pooling (and a strong convolutional baseline). In: Proceedings of the 2018 European Conference on Computer Vision. Cham: Springer, 2018. 480−496
[14]	Su C, Li J N, Zhang S L, Xing J L, Gao W, Tian Q. Pose-driven deep convolutional model for person re-identification. In: Proceedings of the 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 3960−3969
[15]	Sarfraz M S, Schumann A, Eberle A, Stiefelhagen R. A pose-sensitive embedding for person re-identification with expanded cross neighborhood re-ranking. In: Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, Utah, USA: IEEE, 2018. 420−429
[16]	Kalayeh M M, Basaran E, Gökmen M, Kamasak M E, Shah M. Human semantic parsing for person re-identification. In: Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, Utah, USA: IEEE, 2018. 1062−1071
[17]	Gong K, Liang X D, Zhang D Y, Shen X H, Lin L. Look into person: Self-supervised structure-sensitive learning and a new benchmark for human parsing. In: Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Honolulu, Hawaii, USA: IEEE, 2017. 932−940
[18]	Layne R, Hospedales T, Gong S G. Person re-identification by attributes. In: Proceedings of the 23rd British Machine Vision Conference. Guildford, UK: BMVA, 2012.
[19]	Zhu J Q, Liao S C, Lei Z, Li S Z. Multi-label convolutional neural network based pedestrian attribute classification. Image and Vision Computing, 2017, 58: 224−229 doi: 10.1016/j.imavis.2016.07.004
[20]	Deng Y B, Luo P, Loy C C, Tang X O. Pedestrian attribute recognition at far distance. In: Proceedings of the 22nd ACM International Conference on Multimedia. Orlando, Florida, USA: ACM, 2014. 789−792
[21]	Schumann A, Stiefelhagen R. Person re-identification by deep learning attribute-complementary information. In: Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition Workshops. Honolulu, Hawaii, USA: IEEE, 2017. 20−28
[22]	Lin Y T, Zheng L, Zheng Z D, Wu Y, Hu Z L, Yan C G, et al. Improving person re-identification by attribute and identity learning. Pattern Recognition, 2019, 95: 151−161 doi: 10.1016/j.patcog.2019.06.006
[23]	Ristani E, Solera F, Zou R, Cucchiara R, Tomasi C. Performance measures and a data set for multi-target, multi-camera tracking. In: Proceedings of the 2016 European Conference on Computer Vision Workshops. Cham: Springer, 2016. 17−35
[24]	Zheng L, Shen L Y, Tian L, Wang S J, Wang J D, Tian Q. Scalable person re-identification: A benchmark. In: Proceedings of the 2015 IEEE International Conference on Computer Vision. Santiago, Chile: IEEE, 2015. 1116−1124
[25]	Tay C P, Roy S, Yap K H. AANet: Attribute attention network for person re-identifications. In: Proceedings of the 2019 IEEE Conference on Computer Vision and Pattern Recognition. Long Beach, California, USA: IEEE, 2019. 7134−7143
[26]	Zhao X Y, Li H X, Shen X H, Liang X D, Wu Y. A modulation module for multi-task learning with applications in image retrieval. In: Proceedings of the 2018 European Conference on Computer Vision. Cham: Springer, 2018. 401−416
[27]	Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer vision. In: Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016. 2818−2826
[28]	Lin T Y, Goyal P, Girshick R, He K M, Dollar P. Focal loss for dense object detection. In: Proceedings of the 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 2980−2988
[29]	Deng J, Dong W, Socher R, Li L J, Li K, Li F F. Imagenet: A large-scale hierarchical image database. In: Proceedings of the 2009 IEEE Conference on Computer Vision and Pattern Recognition. Miami, Florida, USA: IEEE, 2009. 248−255
[30]	Zhou B L, Khosla A, Lapedriza A, Oliva A, Torralba A. Learning deep features for discriminative localization. In: Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016. 2921−2929
[31]	Selvaraju R R, Cogswell M, Das A, Vedantam R, Parikh D, Batra D. Grad-CAM: Visual explanations from deep networks via gradient-based localization. In: Proceedings of the 2017 IEEE International Conference on Computer Vision. Venice, Italy: IEEE, 2017. 618−626