基于多肤色空间及AdaBoost算法的人脸检测方法

针对复杂背景和可变光照下的彩色图像人脸检测问题,提出一种基于多肤色空间下的肤色分割及Ada-Boost算法的人脸检测方法。首先利用均值滤波、拉普拉斯算子等方法对图像进行增强处理;然后结合YCbCr、YCgCr、YCgCb三种颜色空间下的多肤色空间对图像进行肤色分割,定位出候选的人脸区域;最后对AdaBoost算法的检测过程进行研究验证,检测出人脸并指示。数据显示,该方法在时间、检测率、漏检率等方面都有明显的改进。因此,该方法能较好地处理复杂背景下彩色图像人脸检测的错检、漏检问题,从而提高了检测效率。

Face detection method based on multi-skin color space and AdaBoost algorithm

Focused on face detection problems of color image under the conditions of the complex background and variable illumination, this paper proposed a human face detection method based on multil-skin color space detection and AdaBoost algorithm. First, it used the methods such as the averaging filtering and Laplacian operator to do the image enhancement, and then, combined the three color space of YCbCr, YCgCr, YCgCb to do skin color segmentation and located the candidate's face area. At last, it used AdaBoost algorithm after improving detection process to test and indicate. Statistics show that it has a better detection efficiency, faster speed and lower rate of false.Therefore, the proposed method can handle the problem of false and missing better on face detection of color image under the complex background, and improve efficiency of face detection.