Face Recognition by Combining Kernel Associative Memory and Gabor Transforms

Kernel associative memory (KAM) has previously been proposed as an efficient scheme for face recognition. In this paper, a hybrid method of combining KAM and Gabor wavelet transform is proposed. In this method, face images of each person are first decomposed into their spatial/frequency domains by Gabor transforms, which are then modelled by a KAM. While Gabor properties of orientation selectivity and spatial frequency selectivity provide discriminating features, KAM offers the means to capture the important intra-class variations. Experimental results obtained on two standard face databases demonstrated that the proposed method consistently improved the system performance