一种动态视频流实时多人脸检测算法

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。

A dynamic video stream real-time multiplayer face detection algorithm

With the rapid development of the Internet, researchers are paying more and more attention on the intelligent face detection technology. 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 was proposed in this study. Simulation results show that this method has a lower false detection rate and faster detection speed than traditional methods when used for static images detection and real-time multiple faces detection in dynamic video stream.