一种动态视频流实时多人脸检测算法  

Dynamic Video Stream Real-time Multiplayer Face Detection Algorithm

作　　者：李凯[1] 

机构地区：[1]江西理工大学理学院,江西赣州341000

出　　处：《电视技术》2014年第3期175-177,196,共4页Video Engineering

摘　　要：针对传统的AdaBoost人脸检测算法对侧面及多姿态人脸检测误检率较高且检测速度较慢等问题,提出了一种动态视频流实时多人脸检测算法。仿真实验表明,通过使用该算法对静态图像以及动态视频流中的多个人脸实时检测,其结果比传统算法具有更低的误检率和更快检测速度。Traditional AdaBoost face detection algorithm has problems such as high false detection rate and slow speed when used in side face and multi-pose face detection. A dynamic video stream real-time multiplayer face detection algorithm is proposed in this paper. Simulation results show that this method has a lower false detection rate and faster detection speed than traditional methods when it is used for static images detection and real-time multiple faces detection in dynamic video stream.

关 键 词：动态视频流 ADABOOST 肤色分割 人脸检测 

分 类 号：TN911.73[电子电信—通信与信息系统] TP311[电子电信—信息与通信工程]