Image Fusion Based on Nonsubsampled Contourlet Transform and Pulse Coupled Neural Networks

摘要

In order to overcome the lacking of Shift invariance in Contourlet Transform, enable the image fusion to be in accord with human vision properties, Nonsubsampled Contourlet Transform (NSCT) and Pulse Coupled Neural Networks(PCNN) were used jointly in image fusion algorithms. Original images were decomposed to get the coefficients of low frequency sub bands and high frequency sub bands. The coefficients of low and high frequency sub bands were processed by a modified PCNN. Matching degree of original images is defined and used in fusion rules. Fusion image was obtained by NSCT inverse transformation. Experimental result shows this method is better than Wavelet, Contourlet and traditional PCNN methods, it has bigger mutual information, so the fusion image include more original image's information.