融合肤色分割与人工免疫的人脸检测方法研究

摘要

文章采用YCbCr肤色模型对肤色区域进行判断提取得到肤色候选区域，并对候选区域采用区域填充率方法对明显的非人脸区域进行判断并删除，之后对分割肤色进行二值化处理并采用空间聚类的方法进一步对判断的噪声即非人脸区域进行消除，之后融合人工免疫算法优化肤色区域的聚类算法，提高人脸区域辨识度。%The paper adopt YCbCr elliptic clustering skin color model to extract from the skin color zone, and the candidate region is used to determine the apparent non face region by region iflling rate, After that, the color of the skin is processed and the method of spatial clustering is used to eliminate the noise which is the non face region, ifnally use artiifcial immune algorithm to optimize the color region of the clustering algorithm to improve the recognition of the face region.