DT-CWT Feature Based Classification Using Orthogonal Neighborhood Preserving Projections for Face Recognition

This paper introduces a novel face recognition method based on DT-CWT feature representation using ONPP. The dual-tree complex wavelet transform (DT-CWT) used for representation features of face images, whose kernels are similar to Gabor wavelets, exhibit desirable characteristics of spatial locality and orientation selectivity. And DT-CWT outperforms Gabor with less redundancy and much efficient computing. Orthogonal neighborhood preserving projections (ONPP) is a linear dimensionality reduction technique which attempts to preserve both the intrinsic neighborhood geometry of the data samples and the global geometry. ONPP employs an explicit linear mapping between the two. As a result, ONPP can handle new data samples straightforward, as this amount to a simple linear transformation. The experimental results have demonstrated the advantageous characteristics of ONPP in the DT-CWT feature space and achieve the better face recognition performance