基于YOLO2和ResNet算法的监控视频中的人脸检测与识别  

Face Detection and Recognition in Monitoring Video Based on YOLO2 and ResNet Algorithm

作　　者：朱超平[1,2] 杨艺 ZHU Chaoping;YANG Yi(Chongqing Technology and Business University School of Computer Science and Information Engineering,Chongqing 400067,China;Detection and Control of Integrated Systems Engineering Laboratory,Chongqing 400067,China)

机构地区：[1]重庆工商大学计算机科学与信息工程学院,重庆400067 [2]重庆市检测控制集成系统工程重点实验室,重庆400067

出　　处：《重庆理工大学学报（自然科学）》2018年第8期170-175,共6页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(61502063);重庆工商大学自然科学基金资助项目(1751043)

摘　　要：针对监控环境下的人脸识别比较问题,提出一套基于深度学习的检测识别方法,利用YOLO2算法和Res Net算法完成监控视频中的人脸监测和识别。检测部分采用速度较快的YOLO2算法,识别部分采用准确率较高的Res Net算法。利用WIDER FACE人脸检测数据库和CASIA_Webface数据库做验证性实验,实验结果表明:系统整体的实时性和准确率均可满足实际工程应用需要。A set of detection method based on deep learning is proposed to solve the problem of face recognition in monitoring environment, and our method achieves complete face detection and recognition in monitoring video by using YOLO2 algorithm and Res Net algorithm for the fist time. The detection part takes the fast YOLO2 algorithm,and the recognition part takes Res Net algorithm with high accuracy. Using WIDER FACE monitoring database and CASIA Webface database for verification experiments,the experimental results show that the real time and accuracy of the system can meet the practical application requirements.

关 键 词：深度学习 快速检测 人脸识别 YOLO2算法 ResNet算法 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]