Using phase and magnitude information of the complex directional filter bank for texture segmentation

In this work, a new feature extraction method is proposed for texture segmentation. The approach bases on incorporating the phase information obtained from complex filter banks. The complex directional filter bank (CDFB) is used to decompose a texture image in order to provide complex subband coefficients. The local mean direction, extracted from the phases of the coefficients, is defined as additional features for classification and segmentation. Simulation results show that the CDFB phase information is complementary to the magnitude. Lower classification error rates are achieved. Performance of the proposed method is also compared with other complex filter banks including the Gabor transform and the dual-tree complex wavelet.