In this paper,a series of problems involving Face Recognition were researched in-depth including basic concept of face recognition,background and meaning of face recogniton,histroy of face recogniton,algorithms of face recogniton,application designing based on face recognition,ect.In this paper, we proposed a new face recognition method based on Log-Gabor Filter and KDDA classification. The main advantages of the proposed method :(1) Face representation using Log-Gabor will be more Robust for recognition (2) test four kinds of methods of feature selecting and get the optimizing numbers of selected (3) KDDA reduce feature dimensions and shrink In-Class difference , and also avoid the SSS problem. (4) combine the Log-Gabor filters and Kernel Direct LDA. Recognition experiments with the cas-peal-R1 ,FERET database and some near-infrared face database showed that our method can achieve excellent first one recognition.At last,we developed a picture comparing demo system and designed a Embedded face recogniton system based on above algorithm. |