Skip to Main Content
A recognition algorithm of facial image based on wavelet subbands and decision fusion is discussed in this paper. In this algorithm, wavelet transform is used to decompose facial image into suitable levels of four subbands and then the traditional principal component analysis (PCA) or 2-D Fourier transform (spectroface algorithm) is performed on these four subbands. At last, a decision fusion method is used to get the last result. By comparing with traditional PCA algorithm, PCA algorithm using single wavelet subband and spectroface algorithm, the conclusion is that the algorithm based on wavelet subbands and decision fusion is effective.