摘要
人脸识别是机器视觉和模式识别领域中一个相当困难而又有理论意义和实际价值的研究课题。人脸识别技术通过分析人脸图像,并从人脸图像中提取有效的识别信息,来辨认人的身份。基于特征脸的方法是人脸识别的常用方法。阐述了基于特征脸法的人脸识别的具体步骤算法,再利用ORL人脸库对其进行了测试。试验结果表明,特征脸法一定条件下可以较为准确的识别人脸;而在脸部表情变化较大以及光照条件变化较大的情况下,可以使用加权特征脸法,用人脸原始图像与重构图像的差值来代替原人脸图像而建立新的人脸图像库,取适当的加权系数,能够收到更好的识别效果。
Face recognition is one of the difficult but important issues in the field of machine vision and pattern recognition) which is significant both in theoretical research and practical application.Technology of face recognition is used for identifying identity by analyzing face images and distilling effective recognition information from face images.The method based on eigenface is the common method of face recognition,and the article introduces a standard algorithm of face recognition,then tests with ORL face database,then analyzes the algorithm and proposes the improvement methods.
出处
《长江大学学报（自科版）（上旬）》
CAS
2012年第12期110-112,6,共3页
JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI ＆ ENG
关键词
人脸识别
模式识别
特征脸
ORL
face recognition
pattern recognition
eiginface
ORL
