Optimization for the directional filter bank of the nonsubsampled contourlet transform

The contourlet Transform is an efficient directional multiresolution image representation, but it is not shift-invariant. The nonsubsampled contourlet transform (NSCT), is a fully shift-invariant, multiscale, and multidirection expansion that has a fast implementation, but the computational efficiency is lower than the contourlet transform, such an efficient representation has to be obtained by structured transform and fast algorithm. In this paper, we adopt an optimized directional filter bank and embed it into NSCT to pursue the desired speed, sacrificed the PSNR of the reconstructed image, we obtain the obvious increase of processing speed. Experimental results show that the quality of reconstructed image is sufficient for the human visual system (HVS), and the modified NSCT has a speed about several times than the speed of the customary one, this is very valuable for practical applications of the NSCT.