摘要
基于肤色信息的人脸检测方法能快速获得所有可能的候选人脸,把经过光线补偿处理后的彩色图像从RGB转换到YCbCr色彩空间,根据实验获得的最佳阈值进行肤色分割,快速获得所有可能的候选人脸。对候选人脸作形态学处理及人脸区域标定,最终确定人脸区域。实验结果表明,基于肤色的人脸检测方法能快速的区分人脸区域和背景区域,将人脸区域标定在一个白色的矩形框内,具有良好的检测效果。
Face detection method which base on skin color information can quickly obtain all the possible candidate faces. In this paper, the color images after light compensation are changed from RGB to YCbCr color space. Based on experiment, the best threshold is obtained. Then skin color segmentation is carried out and obtains all the possible candidate faces quickly. Morphologically process candidate faces and locate the face regions, and finally find the face regions. The result of experiment shows that the face detection method which is base on skin color can quickly distinguish face region from background region and can mark the face region into a white rectangular frame, which has excellent detection effect.
出处
《计算技术与自动化》
2008年第4期87-90,共4页
Computing Technology and Automation
关键词
人脸检测
肤色模型
肤色采集
肤色分割
face detection
skin model
skin collection
skin segmentation
