Obtaining Method of Standard Human Face Library in Human Face

摘要

Aiming at static state colorful image, human face was detected by combined complexion division technology with edge detection technology. Firstly, optional face areas were obtained by the area threshold value. Secondly, Gabor transform was employed to locating eyes from the optional face area. Finally, optional face areas were normalization by rotating, scale, and clip processing by using returned position coordinate of two eyes. It can be used for human face identification. Experiments result show that normalized face image is realized by those face image detection and emendation methods, and standard human face library is obtained in the end.