[1]钱华明,王帅帅,王晨宇.基于特征融合的行人重识别算法[J].应用科技,2020,47(2):29-34,43.[doi:10.11991/yykj.201906013]
　QIAN Huaming,WANG Shuaishuai,WANG Chenyu.Research on the person re-identification algorithm based on feature fusion[J].Applied science and technology,2020,47(2):29-34,43.[doi:10.11991/yykj.201906013]
点击复制

基于特征融合的行人重识别算法(/HTML)
分享到：

《应用科技》[ISSN:1009-671X/CN:23-1191/U]

卷:
第47卷
期数:
2020年2期
页码:
29-34,43
栏目:
智能科学与技术
出版日期:
2020-03-05

文章信息/Info

Title:
Research on the person re-identification algorithm based on feature fusion
作者:
钱华明;  王帅帅;  王晨宇
哈尔滨工程大学 自动化学院，黑龙江 哈尔滨 150001
Author(s):
QIAN Huaming;  WANG Shuaishuai;  WANG Chenyu
College of Automation, Harbin Engineering University, Harbin 150001
关键词:
行人重识别; 特征提取; 非线性量化; 颜色命名空间; 直方图; 特征融合; 度量学习; CMC曲线
Keywords:
pedestrian re-identification (Re-ID); feature extraction; nonlineaer quantization; color namespace; histogram; feature fusion; metric learning; CMC curve
分类号:
TP31
DOI:
10.11991/yykj.201906013
文献标志码:
A
摘要:
为解决实际行人重识别系统中识别率低、识别速度慢的问题，从创新和工程应用出发，提出了一种行人重识别算法。对行人图片进行预处理，采用色调、饱和度、亮度（hue，saturation，value，HSV）空间非线性量化的方法构建颜色命名空间，对人体分区域预识别来提高检测效率；对备选目标的整幅图像提取HSV和方向梯度直方图（histogram of oriented gradient，HOG）作为整体特征并在滑动窗口内提取颜色命名（color naming，CN）特征和2个尺度的尺度不变特征（scale invariant local pattern，SILTP），采用本文融合算法得到新的特征；在3个数据集上进行行人重识别，融合的特征在2种度量学习算法的Rank1平均提高了2.4%和3.3%。实验结果表明该算法能够提高重识别精度。
Abstract:
Starting from the innovation and engineering application, a new pedestrian re-identification algorithm was proposed, which mainly solved the problem of low recognition rate and slow recognition speed in the actual pedestrian recognition system. Preprocessing the pedestrian image, using the hue, saturation, and value (HSV) spatial nonlinear quantization method to construct the color namespace, pre-identifying the human sub-regions to improve the recognition speed; extracting the HSV and direction gradient histogram of the entire oriented gradient(HOG) features as the overall feature; and sliding on the entire image of the candidate target, the color naming (CN) features and the scale-invariant local pattern (SILTP) features of the two scales are extracted in the window, getting new features by new fusion algorithms. Pedestrian re-identification is carried out on three data sets. The fusion features improve the average Rank1s of two metric learning algorithms by 2.4% and 3.3% on average. Experimental results show that the algorithm can improve the accuracy of re-identification.

参考文献/References:

[1] 宋婉茹, 赵晴晴, 陈昌红, 等. 行人重识别研究综述[J]. 智能系统学报, 2017, 12(6): 770–780
[2] 张化祥, 刘丽. 行人重识别研究综述[J]. 山东师范大学学报(自然科学版), 2018, 33(4): 379–387
[3] 杨文韬, 李璋, 陈勇, 等. 基于视频图像的行人检测系统的设计与实现[J]. 湖北大学学报(自然科学版), 2017, 39(1): 45–49
[4] FRANCO A, OLIVEIRA L. Convolutional covariance features: Conception, integration and performance in person re-identification[J]. Pattern recognition, 2017, 61: 593–609.
[5] 曾明勇, 吴泽民, 田畅, 等. 基于外观统计特征融合的人体目标再识别[J]. 电子与信息学报, 2014, 36(8): 1844–1851
[6] 张耿宁, 王家宝, 李阳, 等. 基于特征融合与核局部Fisher判别分析的行人重识别[J]. 计算机应用, 2016, 36(9): 2597–2600, 2635
[7] WANG Xuan, LIN Xiaojin, ZHUO Zhiqiang, et al. Pedestrian identification based on fusion of multiple features and multiple classifiers[J]. Neurocomputing, 2016, 188: 151–159.
[8] 刘艳丽, 张纪民. 基于SURF的行人重识别研究[J]. 电脑知识与技术, 2016, 12(22): 167–168
[9] 朱小波, 车进. 基于特征融合与子空间学习的行人重识别算法[J]. 激光与光电子学进展, 2019, 56(2): 021503
[10] WANG Zheng, HU Ruimin, LIANG Chao, et al. Person Re-identification using data-driven metric adaptation[C]//Proceedings of the 21st International Conference on Multimedia Modeling. Sydney, NSW, Australia, 2015: 195–207.
[11] 张见威, 林文钊, 邱隆庆. 基于字典学习和Fisher判别稀疏表不的行人重识别方法[J]. 华南理工大学学报(自然科学版), 2017, 45(7): 55–62
[12] 种衍杰, 方琰, 沙涛. 基于特征融合网络的行人重识别[J]. 计算机系统应用, 2019, 28(1): 127–133
[13] LI Wei, ZHAO Rui, XIAO Tong, et al. DeepReID: deep filter pairing neural network for person re-identification[C]//Proceedings of 2014 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Columbus, OH, USA, 2014: 152–159.
[14] 谭飞刚, 黄玲, 翟聪, 等. 一种用于大型交通枢纽的跨摄像机行人再识别算法研究[J]. 铁道学报, 2017, 39(1): 76–82
[15] ZHAO Rui, OUYANG Wanli, WANG Xiaogang. Person re-identification by saliency learning[J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(2): 356–370.

相似文献/References:

[1]谢红,宁志刚,刘世航.一种新的DK-L人脸识别算法研究[J].应用科技,2009,36(02):42.
　XIE Hong,NING Zhi-gang,LIU Shi-hang.New algorithm of DK-L face recognition[J].Applied science and technology,2009,36(2):42.
[2]陈立伟,倪杰.基于多阈值模糊增强的手指静脉图像分割[J].应用科技,2011,38(04):14.[doi:doi:10.3969/j.issn.1009-671X.2011.04.04]
　CHEN Liwei,NI Jie.Method for finger vein feature segmentation based on multithreshold fuzzy enhancement[J].Applied science and technology,2011,38(2):14.[doi:doi:10.3969/j.issn.1009-671X.2011.04.04]
[3]丁维雷,付永庆.kPCA特征提取算法的自动目标识别[J].应用科技,2011,38(09):32.[doi:10.3969/j.issn.1009-671X.2011.09.08]
　DING Weilei,FU Yongqing.Automatic target recognition based on kPCA feature extraction algorithm[J].Applied science and technology,2011,38(2):32.[doi:10.3969/j.issn.1009-671X.2011.09.08]
[4]张鹏,付希凯,葛国栋,等.二维线性大间距判别分析及其在步态识别中的应用[J].应用科技,2014,41(01):11.[doi:10.3969/j.issn.1009-671X.201306002]
　ZHANG Peng,FU Xikai,GE Guodong,et al.2D linear maximum margin discriminant analysis and its application to gait recognition[J].Applied science and technology,2014,41(2):11.[doi:10.3969/j.issn.1009-671X.201306002]
[5]张春杰,龚再兰,任黎丽.基于修正的Rife和SVM的辐射源特征提取和识别[J].应用科技,2015,42(03):7.[doi:10.3969/j.issn.1009-671X.201403021]
　ZHANG Chunjie,GONG Zailan,REN Lili.Emitter feature extraction and recognition based on the modified Rife and SVM[J].Applied science and technology,2015,42(2):7.[doi:10.3969/j.issn.1009-671X.201403021]
[6]冯驰,胡杨,王兆丰.基于分形理论的涡轮叶片特征提取[J].应用科技,2015,42(04):64.[doi:10.3969/j.issn.1009-671X.201411006]
　FENG Chi,HU Yang,WANG Zhaofeng.Feature extraction of turbine blades based on the fractal theory[J].Applied science and technology,2015,42(2):64.[doi:10.3969/j.issn.1009-671X.201411006]
[7]王素红,宁慧,杨松,等.基于SVM的抄袭检测方法研究[J].应用科技,2015,42(05):51.[doi:10.11991/yykj.201503013]
　WANG Suhong,NING Hui,YANG Song,et al.Research on plagiarism detection method based on SVMs[J].Applied science and technology,2015,42(2):51.[doi:10.11991/yykj.201503013]
[8]马云鹏,李庆武,刘艳,等.基于图像特征融合识别的中文签名鉴伪方法[J].应用科技,2015,42(06):10.[doi:10.11991/yykj.201504011]
　MA Yunpeng,LI Qingwu,LIU Yan,et al.Authenticity identification method for Chinese signature based on image feature fusion recognition[J].Applied science and technology,2015,42(2):10.[doi:10.11991/yykj.201504011]
[9]诸小熊,江加和.基于核相关滤波器的目标跟踪算法[J].应用科技,2017,44(03):48.[doi:10.11991/yykj.201605013]
　ZHU Xiaoxiong,JIANG Jiahe.Visual tracking algorithm based on kernelized correlation filters[J].Applied science and technology,2017,44(2):48.[doi:10.11991/yykj.201605013]
[10]王信,汪友生.基于深度学习与传统机器学习的人脸表情识别综述[J].应用科技,2018,45(01):65.[doi:10.11991/yykj.201707008]
　WANG Xin,WANG Yousheng.Facial expression recognition based on deep learning and traditional machine learning[J].Applied science and technology,2018,45(2):65.[doi:10.11991/yykj.201707008]

备注/Memo

备注/Memo:
收稿日期:2019-06-21。
作者简介:钱华明，男，教授，博士生导师;王帅帅，男，硕士研究生
通讯作者:王帅帅，E-mai：1142846385@qq.com
更新日期/Last Update: 2020-04-21