首先利用人脸的色彩特征和自适应阈值法实现特征候选区域和人脸肤色区域的分离，然后利用人脸的几何特性将连通的特征候选

区域保留下来作为人脸特征区域。后续的特征提取可以在这些人脸特征区域中完成。一般的人脸特征提取方法都可以将该方法作为提高效

率的预处理操作。实验证明，该方法具有高效率、低计算量的特点，并且受人脸表情、图像角度和背景的影响较小。-First, the use of the color characteristics of the human face and adaptive thresholding feature candidate region and the skin color of the face area of ​ ​ separation, then the use of the geometric characteristics of the face to the connectivity features of candidate regions preserved as a facial feature region. Subsequent feature extraction can be completed in the area of ​ ​ these facial feature. General facial feature extraction methods the method as a pre-processing operations to improve efficiency. The experiments show that this method has the characteristics of high efficiency, low computational smaller, and the facial expression, the image angle and background.