信息视频图像人脸检测算法研究

本文首先采用运动信息检测算法,根据帧图像中是否包含运动信息判断图像中是否包含人脸区域,决定是否对该帧进行肤色分割,然后依据非线性的YCbCr肤色模型对需要检测的帧进行分割,进一步确定人脸区域的大致位置,并进行了仿真实验,实验结果表明该算法能够对视频图像序列中的人脸进行检测,具有良好的性能和一定的研究参考价值。

Study of Video-based Face Tracking Techniques

First,The frame is decided whether or not to color segmentation,according to whether the frames contain motion information to determine whether the image contains the face region using motion information.Then the needed for testing frame is divided,according to the YCbCr color model non-linear,and further determined the approximate location of the face region;In this paper,a video-based face tracking system is achieved using MATLAB software.The experimental results show that the system can detect and track face of the video,which has good performance and a certain reference value.