基于肤色模型和模板匹配的人脸检测方法研究

提出了一种基于人脸肤色信息和模板匹配，对彩色图像中的人脸进行检测的方法。该方法首先在YCbCr色彩空间下，利用肤色信息将彩色图像中的肤色区域和非肤色区域分开，计算得到表示肤色相似度的色度图，然后分割由此色度图所生成的灰度图像，从而得到人脸检测的候选区域，同时计算该区域的欧拉数来进一步缩小人脸搜索的范围，结合人脸模板进行匹配，最终确定并标记出彩色图像中人脸的位置。实验结果表明：该方法能够在具有复杂背景的彩色图像中快速并且准确的检测出人脸，误检率较低。

Face Detection Method Based on Skin-color Model and Template

This paper described a method based on skin-color model and template which detecting frontal human faces in color image. In the first step, the method separates the skin regions from non-skin regions in YCbCr chrominance space, and get a chroma chart that shows likelihoods of skin colors. Then we will segment the gray scale image which generate from the original color image to get the candidate region. At the same time, the Euler number of the region is calculated to reduce the face range, matching with template. At last, it identifies and marks the location of the human face. The results of experiments show that the method could be rapid and accurate detection of a human face in the color image with complex back ground, and the rate＇ of false detection is low.