基于视频的人脸检测算法研究

doi:10.16180/j.cnki.issn1007-7820.2020.02.005

摘要/Abstract

摘要：

针对传统AdaBoost算法在视频中检测人脸误检率较高的问题,文中提出了一种结合运动分析和肤色检验的改进型人脸检测算法。该方法通过运动检测来提取运动前景,并选择肤色模型对人脸肤色进行相似度求取,利用几何特征进一步缩小检测范围;采用增加新Haar特征和改进权重更新方式的改进型AdaBoost算法对人脸候选区域进行实时检测。实验结果表明,与传统AdaBoost方法和增加肤色检验的AdaBoost方法相比,该方法的误检率分别降低了18.68%和8.79%,检测时间则分别缩短约800 ms和250 ms。

关键词: 人脸检测, 运动检测, 肤色模型, AdaBoost算法, Haar特征, 权重更新

Abstract:

Aiming at the shortcomings of the traditional AdaBoost algorithm in detecting high false detection rate in video, an improved face detection algorithm combining motion analysis and skin color detection was proposed. The proposed method extracted motion prospects through motion detection, and selected the skin color model to obtain the similarity of the face skin color, and further narrowed the detection range by using the geometric features. The face candidate region was detected in real time detection the improved AdaBoost algorithm which was added with new Haarfeatures and improved weight update mode.The experimental results showed that compared with the traditional AdaBoost method and the AdaBoost method with skin color detection, the false detection rate of this method was reduced by 18.68% and 8.79% respectively, and the detection time was shortened by about 800 ms and 250 ms, respectively.

Key words: face detection, motion analysis, skin color model, AdaBoost algorithm, Haar feature, weight update

中图分类号:

TP391.41

汪欣,吴薇,曾照. 基于视频的人脸检测算法研究[J]. 电子科技, 2020, 33(2): 25-31.

WANG Xin,WU Wei,ZENG Zhao. Research on Face Detection Algorithm Based on Video[J]. Electronic Science and Technology, 2020, 33(2): 25-31.

图/表 12

图1

图2

图3

图4

图5

图6

图7

图8

图9

图10

表1

表2

参考文献 15

[1]	崔鹏, 燕天天 . 融合YCbCr肤色模型与改进的Adaboost算法的人脸检测[J]. 哈尔滨理工大学学报, 2018,23(2):91-96.
	Cui Peng, Yan Tiantian . Face detection based on YCbCr skin model and improved adaboost algorithm[J]. Journal of Harbin University of Science and Technology, 2018,23(2):91-96.
[2]	糜元根, 陈丹驰, 季鹏 . 基于几何特征与新Haar特征的人脸检测算法[J]. 传感器与微系统, 2017,36(2):154-157.
	Zhai Yuangen, Chen Danchi, Ji Peng . Face detection algorithm based on geometric features and new Haar features[J]. Journal of Sensors and Microsystems, 2017,36(2):154-157.
[3]	Rajeshwari J, Karibasappa K . Adaboost modular tensor locality preservative projection:face detection in video using AdaBoost modular-based tensor locality preservative projections[J]. IET Computer Vision, 2017,10(7):670-678.
[4]	闵秋莎, 刘能, 陈雅婷 , 等. 基于面部特征点定位的头部姿态估计[J]. 计算机工程, 2018,48(6):269-275.
	Qiu Qisha, Liu Neng, Chen Yating , et al. Head pose estimation based on facial feature point location[J]. Computer Engineering, 2018,48(6):269-275.
[5]	赵喆, 侯俊 . 基于肤色模型和改进Adaboost算法的人脸检测[J]. 电子科技, 2015,28(12):80-83.
	Zhao Zhe, Hou Jun . Face detection based on skin color model and improved Adaboostalgorithm[J]. Electronic Science and Technology, 2015,28(12):80-83.
[6]	Zhao B . Target monitoring on face detection based on improved AdaBoost algorithm[J]. Lecture Notes in Electrical Engineering, 2014,30(18):373-377.
[7]	陈章乐, 蔡茂国, 刘凡秀 . 一种结合肤色及类人脸特征的人脸检测[J]. 计算机工程与应用, 2013,49(3):194-197.
	Chen Zhangle, Cai Maoguo, Liu Fanxiu . Face detection based on skin color and face-like features[J]. Computer Engineering and Applications, 2013,49(3):194-197.
[8]	Tavallali P, Yazdi M, Khosravi M R . Robust cascaded skin detector based on AdaBoost[J]. Multimedia Tools & Applications, 2018,26(4):1-22.
[9]	李扬, 曹著明 . 基于肤色和AdaBoost算法的人脸检测[J]. 现代计算机, 2017,35(5):79-82.
	Li Yang, Cao Zhuming . Face detection based on skin color and AdaBoost algorithm[J]. Modern Computer, 2017,35(5):79-82.
[10]	方承志, 袁海峰 . 基于CMYK-H-CbCr肤色检测和改进型AdaBoost算法的人脸检测[J]. 计算机应用与软件, 2017,34(8):167-172.
	Fang Chengzhi, Yuan Haifeng . Face detection based on CMYK-H-CbCr skin color detection and improved AdaBoost algorithm[J]. Journal of Computer Applications and Software, 2017,34(8):167-172.
[11]	Lv S D, Song Y D, Xu M , et al. Face detection under complex background and illumination[J]. Journal of Electronic Science and Technology, 2015,13(1):78-82.
[12]	罗富丽, 李佳田 . 结合运动特征的AdaBoost层次增强算法[J]. 计算机工程与应用, 2017,53(7):154-159.
	Luo Fuli, Li Jiatian . AdaBoost hierarchy enhancement algorithm based on motion characteristics[J]. Computer Engineering and Applications, 2017,53(7):154-159.
[13]	Sun B, Chen S, Wang J , et al. A robust multi-class AdaBoost algorithm for mislabeled noisy data[J]. Knowledge-Based Systems, 2016,102(5):87-102.
[14]	周振华 . AdaBoost人脸检测定点型优化算法[J]. 电子科技大学学报, 2015,24(4):589-593.
	Zhou Zhenhua . AdaBoost face detection and fixed point optimization algorithm[J]. Journal of University of Electronic Science and Technology of China, 2015,24(4):589-593.
[15]	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.

检测算法	实际人脸个数	检测人脸数	正检人脸数	错检人脸数	漏检率/%	误检率/%
传统AdaBoost	204	253	202	51	0.98	20.15
改进AdaBoost	204	224	201	23	1.47	10.26
本文算法	204	203	200	3	1.96	1.47

检测算法	视频格式
检测算法	1 280×1 080	1 280×960
传统的AdaBoost	983	805
改进的AdaBoost	358	312
本文算法	92	86