摘要
人眼检测对于人脸的自动识别非常重要。本文提出了一种视频图像序列中快速人眼定位的方法。确定一个眼睛窗口,综合利用窗口内灰度的均值、灰度方差的均值、梯度的均值并结合色度信息来进行眼睛的检测。将提出的眼睛定位的方法应用到视频会议中,从多个摄像机得到的图像中,根据一个判断函数判断出正面人脸图像。实验结果表明,所提算法能够迅速准确的检测出眼睛并且能够正确的判断出正面人脸图像。
Eye detection is very important for automatic face recognition. A method for fast eye location in a sequence of visual pictures was proposed in this paper. Firstly, an eye window was determined. Then, eye was detected according to the expectation of gray levels, the expectation of the variance of gray levels, the expectation of gradients of the gray levels and chroma information in the window. The proposed eye - positioning method was applied to the video conference and according to a judgment function, the obverse human face image was determined from the images captured by some cameras. The experiment results demonstrate that the proposed algorithm can detect the eyes rapidly and precisely and judge the obverse human face image accurately.
出处
《北京电子科技学院学报》
2008年第2期49-51,58,共4页
Journal of Beijing Electronic Science And Technology Institute
关键词
人脸检测
人眼检测
模式识别
灰度投影
face detection
eye detection
pattern recognition
gray projection
