多人脸跟踪与最佳人脸提取

doi:10.16180/j.cnki.issn1007-7820.2019.09.007

摘要/Abstract

摘要：

针对视频人脸识别系统中同一人脸重复识别的问题,文中提出了一种多人脸跟踪与最佳人脸提取的方法。通过ViBe算法提取运动区域,缩小数据处理区域及确定执行人脸检测;利用Haar特征结合AdaBoost算法检测人脸,并根据肤色检测判断是否有误检;利用CamShift算法跟踪人脸;再使用Sobel算子得到清晰的人脸图片。实验表明,该方法下人脸误检率由2.8%降到0.2%,对于100帧视频平均处理时间从原始每帧112 ms降低到了45.6 ms,其处理速度明显提升。

关键词: 人脸检测, 人脸跟踪, AdaBoost, Haar, CamShift, Sobel

Abstract: Aim

ing at the problem that repeated recognition of the same person in video face recognition system, a multi-face tracking and optimal face extraction method was proposed. Using ViBe modeling to extract the motion area, reducing the data processing area; The Haar feature was combined with the AdaBoost algorithm to detect the face, and the skin color detection was used to determine whether there was a false detection; Tracking faces with CamShift algorithm; Using the Sobel operator to got a clearer face image. Experiments showed that under this method, the face false detection rate was reduced from 2.8% to 0.2%. For 100 frames, the average processing time was reduced from 112 milliseconds per frame to 45.6 milliseconds, and the processing speed was significantly improved.

Key words: face detection, face tracking, AdaBoost, Haar, CamShift, Sobel

中图分类号:

TP391.41

田雄,吴薇,刘晓尚,吴秀. 多人脸跟踪与最佳人脸提取[J]. 电子科技, 2019, 32(9): 32-37.

TIAN Xiong,WU Wei,LIU Xiaoshang,WU Xiu. Multi-face Tracking and Optimal Face Extraction[J]. Electronic Science and Technology, 2019, 32(9): 32-37.

图/表 9

图1

图2

图3

图4

图5

图6

表1

图7

图8

参考文献 18

[1]	Exner D, Bruns E, Kurz D, et al. Fast and robust CamShift tracking [C].Seattle:Computer Vision and Pattern Recognition Workshops,IEEE, 2010.
[2]	汤泉 . 基于AdaBoost和CamShift的人脸检测与跟踪[J]. 电子科技, 2016,29(12):166-169.
	Tang Quan . Face detection and tracking based on AdaBoost and CamShift[J]. Electronic Science and Technology, 2016,29(12):166-169.
[3]	慕春雷 . 基于HOG特征的人脸识别系统研究[D]. 成都:电子科技大学, 2013.
	Mu Chunlei . Research on face recognition system based on HOG feature[D]. Chengdu:University of Electronic Science and Technology of China, 2013.
[4]	李天成, 范红旗, 孙树栋 . 粒子滤波理论、方法及其在多目标跟踪中的应用[J]. 自动化学报, 2015,41(12):1981-2002. doi: 10.16383/j.aas.2015.c150426
	Li Tiancheng, Fan Hongqi, Sun Shudong . Particle filter theory and method and its application to multi-target tracking[J]. Journal of Automation, 2015,41(12):1981-2002. doi: 10.16383/j.aas.2015.c150426
[5]	张明杰, 康宝生 . 一种基于图模型的粒子滤波跟踪方法[J]. 计算机应用研究, 2016,33(2):590-593.
	Zhang Mingjie, Kang Baosheng . A particle filter tracking method based on graph model[J]. Journal of Computer Applications, 2016,33(2):590-593.
[6]	闫钧华, 陈少华, 艾淑芳 , 等. 基于Kalman预测器的改进的CamShift目标跟踪[J].中国惯性技术学报, 2014(4):536-542.
	Yan Junhua, Chen Shaohua, Ai Shufang , et al. An improved CamShift target tracking based on Kalman predictor[J].Journal of Chinese Inertial Technology, 2014(4):536-542.
[7]	王丽, 郝晓丽 . 基于Kalman滤波器和改进Camshift算法的双眼跟踪[J]. 微电子学与计算机, 2016,33(6):109-112.
	Wang Li, Hao Xiaoli . Binocular tracking based on Kalman filter and improved CamShift algorithm[J]. Microelectronics & Computer, 2016,33(6):109-112.
[8]	甘玲, 朱江, 苗东 . 扩展Haar特征检测人眼的方法[J]. 电子科技大学学报, 2010,39(2):247-250.
	Gan Ling, Zhu Jiang, Miao Dong . Method for detecting human eye by extending Haar feature[J]. Journal of University of Electronic Science and Technology of China, 2010,39(2):247-250.
[9]	王庆伟, 应自炉 . 一种基于Haar-Like T特征的人脸检测算法[J]. 模式识别与人工智能, 2015,28(1):35-41. doi: 10.16451/j.cnki.issn1003-6059.201501005
	Wang Qingwei, Ying Zilu . A face detection algorithm based on Haar-Like T feature[J]. Pattern Recognition & Artificial Intelligence, 2015,28(1):35-41. doi: 10.16451/j.cnki.issn1003-6059.201501005
[10]	马健 . 结构化Haar特征方法研究与应用[D]. 西安:西安电子科技大学, 2014.
	Ma Jian . Research and application of structured Haar feature method[D]. Xi’an: Xidian University, 2014.
[11]	陈锻生, 刘政凯 . 肤色检测技术综述[J]. 计算机学报, 2006,29(2):194-207.
	Chen Chengsheng, Liu Zhengkai . A review of skin color detection technology[J]. Chinese Journal of Computers, 2006,29(2):194-207.
[12]	史东承, 倪康 .基于 YCbCr 颜色空间背景建模与手势阴影消除[J]. 中国光学, 2015,8(4):589-595.
	Shi Dongcheng, Ni Kang . Modeling and gesture shadow elimination based on YCbCr color space[J]. China Optics, 2015,8(4):589-595.
[13]	樊慧慧, 王金海 . 一种复杂环境下的动态多人脸自动检测与实时跟踪方法[J]. 计算机应用研究, 2015,32(10):3183-3186.
	Fan Huihui, Wang Jinhai . An automatic multi-face detection and real-time tracking method in complex environment[J]. Application Research of Computers, 2015,32(10):3183-3186.
[14]	李超, 刘铁根, 刘宏利 , 等. 基于Haar检测和改进的CamShift的人脸跟踪[J].光电子·激光, 2011(12):1852-1856.
	Li Chao, Liu Tiegen, Liu Hongli , et al. CamShift face detection based on Haar detection and improvement[J].Journal of Optoelectronics Laser, 2011(12):1852-1856.
[15]	赵勇, 李怀宇 . 基于改进Sobel算法的车辆检测技术[J]. 电子科技, 2017,30(11):78-80.
	Zhao Yong, Li Huaiyu . Vehicle detection technique based on improved sobel algorithm[J]. Electronic Science and Technology, 2017 ( 11):78-80.
[16]	孟繁静, 王连明 . 视频人脸识别系统中的最佳人脸捕获[J]. 东北师大学报:自然科学版, 2015,47(3):90-95.
	Meng Fanjing, Wang Lianming . The best face capture in video face fecognition system[J]. Journal of Northeast Normal University:Natural Science Edition, 2015,47(3):90-95.
[17]	沈铮, 吴薇 . 基于视频图像的公交车人群异常情况检测[J]. 南京理工大学学报:自然科学版, 2017,41(1):65-73.
	Shen Zhen, Wu Wei . Video-based abnormal crowd behavior detection on bus[J]. Journal of Nanjing University of Science and Technology:Natural Science, 2017,41(1):65-73.
[18]	Barnich O, Droogenbroeck M V. ViBe:A powerful random technique to estimate the background in video sequences [C]. Cambridge:IEEE International Conference on Acoustics, Speech and Signal Processing,IEEE, 2009.

算法	Corr nums	Error nums	Refuse detect	Corr rates	Error rates	Refuse rates
1	463	14	23	92.6%	2.8%	4.6%
2	476	1	23	95.2%	0.2%	4.6%