Multimodal medical image fusion based on nonsubsampled contourlet transform using improved PCNN

Multimodal medical image fusion is an indispensable branch in the field of image fusion. In order to obtain a more complete and more reliable medical image, this paper presents a novel approach for multimodal medical image fusion using an improved pulse-coupled neural network (IPCNN) in nonsubsampled contourlet transform (NSCT) domain. First, the image is decomposed into sub-bands with different scales and different directions by NSP and NSDFB. Next, local area singular value is introduced to determine the structural information factor which will be the linking strength parameter of PCNN. After the fire process we can get the fire mapping images that can reflect the characteristics of single pixel and its neighborhood. Then, we extract the objects with salient features of the fire mapping images by compare-selection operator. Finally, we construct the fused image by inverse NSCT. Our proposed algorithm in multimodal medical image fusion is proved to perform better in robustness and reliability over the existing methods, meeting the requirement of human vision.