基于肤色模型的人脸检测与人眼定位

Face Detection and Eyes Localization Based on YCbCr Skin Model

针对光照变化对人脸检测及人眼定位的影响,提出一种基于肤色模型的人脸检测与人眼定位方法.先对图像进行预处理,减少图像中的噪声;再将RGB颜色空间转化到具有良好肤色聚类特性的YCbCr色彩空间,利用Gauss模型进行肤色建模;最后检测出人脸区域并确定人眼位置.

In view of illumination change, a method of face detection and YCbCr skin model was proposed. First, in order to reduce the noise in eyes localization based on the image, the image was preprocessed; then, RGB color space was transformed into YCbCr color space with good skin color clustering characteristics, and skin model was built through the method of Gaussian model. Finally, the face area was detected and the eye position was determined.