抄録
The proposed scheme based on holistic information of face image which is obtained by DCT and multi-stage wavelet analysis and 2D-PCA classifier to classify the facial feature to a person’s class. The objectives of the proposed method are to find the best frequency-based features extraction when it is combined with 2D_PCA, and to reduce the high space requirements of classical PCA. The facial feature is built by keeping small part of frequency domain coefficients, which have big magnitude value. Next, the facial feature is analyzed using 2D-PCA for finding the class separation. From the experimental results can be concluded that the frequency analysis is an efficient way to reduce memory space requirements and computational load of classical PCA and 2D-PCA. The DCT-based facial feature gives the best performance.
