This paper presents a new feature extraction technique for recognizing human faces. The proposed method uses hybrid feature extraction techniques such as Principal Component Analysis and Gabor Wavelet are combined together. The classifier k-Nearest Neighbor is used for classification. For experimentation JAFEE and YALE face databases are used to test and achieved better performance of proposed method. The performance of proposed method on JAFFE and Yale face database are compared with known other methods.
Keywords
PCA, Wavelet, k-NN, JAFEE database, Yale face databases.