基于肤色模型与模板匹配的人脸检测探究

摘要

本文提出了一种基于肤色模型与模板匹配的人脸检测方法。基于YCbCr色彩空间的高斯概率模型，对肤色进行相似度计算和二值化分割，采用形态学对二值化后的图像去噪处理，随后又利用人脸先验知识进一步排除非人脸区域，最后结合人脸模板进行匹配，最终确定并标记出图像中人脸的位置。实验表明，该方法能够在复杂背景图像中较准确地检测出人脸。%The face detection method based on skin color model and template matching is presented in this paper.With the skin color Gaussin probability model based on YCbCr color space,the color similarity is calculated and two value segmentation is processed.The morphological filter is used for two value of the image.Subsequently,the face’s priori knowledge is exploited to further exclude non-human face area. Finally,the use of face template matching,eventually identifies and marks out face location of the human image.The results of experiments show that the method could be accurate detection of a human face in the image with complex back ground.