机构地区: 华南农业大学信息学院
出 处: 《现代计算机》 2009年第6期60-62,共3页
摘 要: 讨论并实现一个基于肤色模型和CAMShift方法的人脸检测与跟踪原型系统。该系统采用肤色模型分割出视频帧中的肤色区域与非肤色区域以检测出人脸,利用CAMShift算法跟踪运动的人脸,完成对人脸各种姿态的跟踪,具有较好的实时性和鲁棒性。 Discusses a method for human face detection and tracking based on skin-color model and continuously adaptive mean shift algorithm and realizes a corresponding prototype. This method uses skin-color model to separate skin-tone regions from non-face regions in the video frame, and then uses the CAM Shift algorithm tracking the face detected before. The experiment shows that it can robustly track human faces with different poses in real-time.