基于梯度信息的灰度人脸检测

针对灰度图像提出了一种有效的人脸检测算法。首先对原始灰度图像利用Sobel边缘检测算子得到水平方向的梯度信息;在此基础上,结合人的眼睛灰度信息得到眼睛的初步候选区域,并根据眼睛的几何特征删除非眼睛区域;然后利用眼睛的几何特征以及人脸的“三停五眼”特性确定真正的眼睛和嘴的区域,在精确定位眼睛和嘴的几何位置后确定人脸区域;最后利用人脸的左右对称性来验证人脸。实验表明,该算法对于灰度图像的正面人脸定位以及人脸检测具有良好的效果。

Gray Face Detection Based On Grads Information

Gray face detection plays an important role in face recognition,the current detection,however,are mostly based on color images,and it is hard to see face detection based on gray images.An effective and rapid face detection method for gray images is proposed in the present paper.First of all,the horizontal grads information of original image containing faces are obtained by means of Sobel operator,and these information are then associated with human eye's gray level information so that the candidate regions of possible eyes can be obtained.After that,the regions of false eyes are eliminated by taking such geometric features of human eyes as symmetry and shape into account,and according to the concept that a standard face could be partitioned equally into three parts vertically and five parts horizontally,the real regions of eyes and mouths could then be obtained.The possible face regions can be located based on the corresponding geometric positions of eyes and mouths.At last,their symmetries can be calculated and the possible face can be verified from the possible existing facial features.Experiments show that this algorithm achieves satisfactory performance in face location and face detection.