EFFICIENT FUSION OF MULTIMODAL MEDICAL IMAGES USING NON-SUB SAMPLED CONTOUR LET TRANSFORM

摘要

Image fusion for the multimodal images would provide wide applications in the field of medical sciences. The main motivation is to capture the relevant information from the medical image sources and fuse them together to provide a single output which forms as an important system in the medical diagnosis. In this paper a fusion framework is provided for the multimodal medical image fusion using non-subsampled contourlet transform (NSCT). The NSCT uses two kinds of methodology for image fusion, the phase congruency and directive contrast technique. The former is used to fuse the low frequency coefficients of the images whereas the latter is used for the high frequency coefficients of the images. Various multimodal medical images are given as input where the images are decomposed and finally the fused images are constructed by using the inverse NSCT technique. The proposed framework will provide a way to enable much more accurate analysis of the fused images. The applicability of the proposed idea can be carried out with three clinical examples like brain affected with Alzheimer disease, subaccute stroke, brain tumour.