Face Recognition Using Wavelet Transform and Locally Discriminating Projection

In this paper, wavelet based partial representation of a face is used along with locally discriminating projection (LDP) for automatic face recognition task. The face image is decomposed into different subbands and the subbands which contain the discriminatory information for face recognition are selected for face representation. A novel, local structure based supervised feature extraction method based on the idea of Locality Preserving Projection (LPP) is proposed along with wavelet features to enhance the class structure of the data with local and directional information. Experimental results on the Yale and ORL face databases suggest that the proposed approach provides a better representation and achieves lower error rates in face recognition.