摘要
人脸识别系统从广义上说大致可以分为人脸检测、定位与跟踪,人脸表征,人脸识别,表情/姿势分析,生理分类等五个方面.人脸的特征提取和人脸图像匹配这两部分是最重要的,也是决定人脸识别系统的准确性和效率所在.采用2D-DCT变换进行人脸特征提取和EHMM模型进行人脸图像匹配(人脸识别),并用多样本加权合成的EHMM训练的方法,取得了较高的速度和识别率.
Generally speaking,face recognition system is made up of five groups,which are called face detection, location and tracing, face representation, face recognition, expression/gesture analysis and face physical classification. In the face recognition systerm, face detection and face recognition are the most important facets. They also determine the face recognition system＇s correction and efficiency. In this paper, the 2D-DCT in face eigenvalue＇s extraction and EHMM mode in face image matching （face recognition ）, and the multiswatch weight synthetic trained method are explayed. It is obvious that the method has gained better recognition rate and rapidity.
出处
《云南民族大学学报（自然科学版）》
CAS
2006年第4期285-288,共4页
Journal of Yunnan Minzu University:Natural Sciences Edition
