基于特征聚类对群三元组损失的车辆再识别

doi:10.3969/j.issn.0372-2112.2020.12.021

电子学报 ›› 2020, Vol. 48 ›› Issue (12): 2444-2452.DOI: 10.3969/j.issn.0372-2112.2020.12.021

基于特征聚类对群三元组损失的车辆再识别

吴燕雄^1,2, 蔡建羡¹, 滕云田^1,2

1. 防灾科技学院电子科学与控制工程学院, 河北三河 065201;
2. 中国地震局地球物理研究所, 北京 100081

收稿日期:2020-04-10 修回日期:2020-05-11 出版日期:2020-12-25
- 通讯作者:
- 滕云田
- 作者简介:
- 吴燕雄 女,1983年出生,辽宁沈阳人,于2006年和2009年分别获得北京航空航天大学学士学位和硕士学位.现为中国地震局防灾科技学院讲师,并为中国地震局地球物理研究所博士生,主要研究方向为智能信息处理、智能传感器和观测技术.E-mail:wuyanxiong@cidp.edu.cn;蔡建羡 女,1978年出生,河北衡水人,2001年于河北科技大学获学士学位,2003年于燕山大学获硕士学位,2010年于北京工业大学获博士学位.现为防灾科技学院学院教授,主要研究方向为:机器学习、地震仪器监测技术、智能机器人技术.E-mail:cjxlaq@163.com
- 基金资助:
- 防灾科技学院教学研究与教育改革项目 (No.JY2016B10）; 河北省高等学校科学技术研究重点项目 (No.ZD2018304）; 中央高校基本科研业务费 (No.ZY20180111）; 国家重点研发计划项目 (No.2018YFC1503801）

Vehicle Re-identification Using the Coupled Feature Clusters Embedded into Triplet Loss

WU Yan-xiong^1,2, CAI Jian-xian¹, TENG Yun-tian^1,2

1. School of Electronic Science and Control Engineering, Institute of Disaster Prevention, Sanhe, Hebei 065201, China;
2. Institute of Geophysics, China Earthquake Administration, Beijing 100081, China

Received:2020-04-10 Revised:2020-05-11 Online:2020-12-25 Published:2020-12-25
- Corresponding author:
- TENG Yun-tian
- Supported by:
- Teaching Research and Education Reform Project of Institute of Disaster Prevention (No.JY2016B10); Key Project of Science and Technology Research in Higher Education of Hebei Province (No.ZD2018304); Fundamental Research Funds for the Central Universities (No.ZY20180111); National Key Research and Development Program of China (No.2018YFC1503801)

摘要/Abstract

摘要： 车辆再识别旨在从多个摄像机拍摄的图像中识别出同一车辆.本文提出了一种对群三元组损失函数，以特征中心点替代均值，并将对群思想和三元组损失相结合，优化了困难样本的识别.车辆再识别过程中，对群损失函数的训练过程扩大了样本规模，增加了计算量，且传统对群损失函数无法准确处理困难正样本.为此，提出了一种特征聚类对群三元组损失函数.本方法采用正样本特征聚类中心并改进了三元组损失函数的设计，从而优化了对群损失函数.在不扩增输入样本数量的同时提升了算法处理困难样本的能力.实验表明，与主流车辆再识别算法相比，本方法可有效提升车辆再识别的准确率.

关键词: 车辆再识别, 视觉特征, 特征聚类对群损失, 三元组损失

Abstract: Vehicle re-identification is the task of identifying the same vehicle across some images captured by multiple cameras. We propose a coupled feature clusters embedded into triplet loss dealing with hard samples. During the vehicle re-identification, the coupled clusters loss suffers from larger computation consumption caused by the extension of the sample scale and the reduction of identification accuracy. Therefore, the coupled feature clusters embedded into triplet loss is proposed. It improves the ability of the algorithm on processing hard samples in terms of selecting feature centers of positive samples based on clustering and the embedded into a triple loss. Experiments show that the algorithm effectively improves the accuracy of vehicle re-identification compared to the vehicle re-identification algorithm based on coupled clusters loss.

Key words: vehicle re-identification, visual appearance, coupled feature clusters loss, triple loss

中图分类号:

TP391.4

吴燕雄, 蔡建羡, 滕云田. 基于特征聚类对群三元组损失的车辆再识别[J]. 电子学报, 2020, 48(12): 2444-2452.

WU Yan-xiong, CAI Jian-xian, TENG Yun-tian . Vehicle Re-identification Using the Coupled Feature Clusters Embedded into Triplet Loss[J]. Acta Electronica Sinica, 2020, 48(12): 2444-2452.

参考文献

[1] Guo H,Zhao C,Liu Z,et al.Learning coarse-to-fine structured feature embedding for vehicle re-identification[A].32th AAAI Conference on Artificial Intelligence[C].New Orleans,Louisiana,USA:AAAI,2018.6853-6860.
[2] Shen Y,Xiao T,Li H,et al.Learning deep neural networks for vehicle Re-ID with visual-spatio-temporal path proposals[A].2017 IEEE International Conference on Computer Vision[C].Venice,Italy:ICCV,2017.1900-1909.
[3] Zapletal D,Herout A.Vehicle re-identification for automatic video traffic surveillance[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].Las Vegas,Nevada,USA:IEEE,2016.25-31.
[4] Liu X,Liu W,Ma H,et al.Large-scale vehicle re-identification in urban surveillance videos[A].IEEE International Conference on Multimedia and Expo[C].Seattle,WA,USA:IEEE,2016.1-6.
[5] Liu H,Tian Y,Wang Y,et al.Deep relative distance learning:tell the difference between similar vehicles[A].Computer Vision and Pattern Recognition[C].Las Vegas,Nevada,USA:IEEE,2016.2167-2175.
[6] Zhou Y,Shao L.Viewpoint-aware attentive multi-view inference for vehicle re-identification:computer vision and pattern recognition[A].Computer Vision and Pattern Recognition[C].Salt Lake City,UT,USA:IEEE,2018.6489-6498.
[7] Wang Z,Wang X,Tang L,et al.Orientation invariant feature embedding and spatial temporal regularization for vehicle re-identification[A].2017 IEEE International Conference on Computer Vision[C].Venice,Italy:IEEE,2017.379-387.
[8] Xiao T,Li H,Ouyang W,et al.Learning deep feature representations with domain guided dropout for person re-identification[A].IEEE Conference on Computer Vision and Pattern Recognition[C].Las Vegas,Nevada,USA:IEEE,2016.1249-1258.
[9] Liao S,Hu Y,Zhu X,et al.Person re-identification by local maximal occurrence representation and metric learning[A].IEEE Conference on Computer Vision and Pattern Recognition[C].Boston,Massachusetts,USA:IEEE,2015.2197-2206.
[10] Wu C W,Liu C T,Chiang C E,et al.Vehicle re-identification with the space-time prior[A].In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops[C].Salt Lake City,UT,USA:IEEE,2018.121-128.
[11] Chen G,Lu J,Yang M,et al.Spatial-temporal attention-aware learning for video-based person re-identification[J].IEEE Transactions on Image Processing,2019,28(9):4192-4205.
[12] He B,Li J,Zhao Y,et al.Part-regularized near-duplicate vehicle re-identification[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].Long Beach,CA,USA:IEEE,2019.3997-4005.
[13] Liu X,Zhang S,Huang Q,et al.Ram:a region-aware deep model for vehicle re-identification[A].2018 IEEE International Conference on Multimedia and Expo (ICME)[C].San Diego,USA:IEEE,2018.1-6.
[14] Hermans A,Beyer L,Leibe B.In Defense of the Triplet Loss for Person Re-identification[EB/OL].https://arxiv.org/abs/1703.07737,2017.
[15] Yadav J,Sharma M.A review of K-mean algorithm[J].International journal of engineering trends and technology,2013,4(7):2972-2976.
[16] He K,Zhang X,Ren S,Sun J.Deep residual learning for image recognition[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].Las Vegas,Nevada,USA:IEEE,2016.770-778.
[17] Liu X,Liu W,Mei T,Ma H.A deep learning-based approach to progressive vehicle re-identification for urbansurveillance[A].ECCV 2016[C].Amsterdam,Netherlands:Springer, 2016.869-884.

基于特征聚类对群三元组损失的车辆再识别

Vehicle Re-identification Using the Coupled Feature Clusters Embedded into Triplet Loss

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

[1]	戴文战, 姜晓丽, 李俊峰. 基于人眼视觉特性的NSCT医学图像自适应融合[J]. 电子学报, 2016, 44(8): 1932-1939.
[2]	管业鹏. 基于多模态视觉特征的自然人机交互[J]. 电子学报, 2013, 41(11): 2223-2229.