摘要
提出一种有效的在复杂背景和不同光照条件下从彩色图像中检测出人脸的方法。该方法由两个阶段组成:第一阶段利用彩色信息和基于等腰三角形分割的方法找到可能的人脸区域。第二阶段利用二维相关系数作为对称相似性测度进行人脸认证。提出的方法能够处理各种尺寸、不同光照条件、不同的姿态及表情变化的人脸,在复杂背景的彩色图像中提高了人脸的检测速度。实验结果显示该方法获得良好的效果。
Develops an efficient face detection scheme that can detect faces in color images with complex environments and different illumination levels, The proposed scheme comprises two stages. In the first stage color and triangle- based segmentation is adapted to search potential face ragions. The second stage involves the 2 -- D correlation coefficient used as a symmetry similarity mearsttre for face verification. It can handle various sizes of faces, different illumination conditions, divers looses and changeable expression; and increase the execution speed of the face detection algorithm in the case of complex backgrounds. Results of experiment demonstrate that method obtains good results.
出处
《计算机技术与发展》
2009年第2期52-54,57,共4页
Computer Technology and Development
基金
国家自然科学基金(60572128)
