A novel algorithm for solving the small sample size problem in face recognition is proposed. The size of each face sample is increased greatly and the small sample size problem is resolved by regarding every Gabor wavelet transformed output image as an independent sample. A special nearest neighbor classifier using whitened cosine distance is designed for the classification of face feature vector groups. The experiments comparing the prosposed approach with the direct PCA method in FERET face database show that the average correct face recognition rate of the proposed algorithm can reach 97%

which is superior to the direct PCA method.