视频图像检测中嘴部错检的自动更正算法

提出一种应用于视频序列图像的新的嘴部检测算法．该算法能克服现有的嘴部检测算法的缺陷，利用视频中相邻帧图像之间的高度相关性，能探测到现有算法中的误检、漏检，并自动作出纠正，得到正确的人嘴检测结果。实验结果显示，该算法能大大提高视频序列图像中嘴部检测的正确率。

Self-Correction Algorithm for Mouth Failure Detection in Video Image Detection

Puts forward a kind of new mouth detection algorithm applied in video image sequence, and the algorithm can overcome the defect of existing mouth detection algorithm. Uses the high correla- tion between adjacent frames of image in video, and it can detect the error, omission of the ex- isting algorithm, and automatically corrects, so as to get the right people nozzle testing results. The experimental result shows that, the algorithm can greatly improve the video image sequence mouth detection accuracy.