Multi-Modality Medical Image Fusion Based On Contourlet Transform And Fuzzy Logic

This research mainly focuses on Medical Image Fusion,because the successful diagnosis of any disease,retrieval of images,undergoes surgery treatment,tumor identification etc.depends on the accuracy of the image obtained from medical imaging modalities.Several schemes and algorithm are done and many are proposed for image fusion but it is impossible to devise a fixed universal method valid to all image fusion tasks.In this research,we proposed and presented an image fusion algorithms for the fusion of multi-modality medical images(such as Computed Tomography(CT)which gives anatomical structure of hard tissues like bones,and Magnetic Resonance Imaging(MRI)which deals the anatomical structure of soft tissues,organs and blood vessels),based on Contourlet Transform and Fuzzy Reasoning.First two(CT and MRI)images of the same scene are selected from any registered source.Contourlet Transform according to the lat est studies is consider more powerful tool for analysis of the 2-D signals(images)contain lines,curves,edges or hyper-plane singularity than wavelet and other MGA tools,is applied on both the source images separately.It decomposes both the CT and MRI images into High Frequency Subbands(HFS)coefficients and Low Frequency Subbands(LF S)coefficients.In second stage the Fuzzy Logic which is an efficient intelligent,soft computing method to handle uncertain information and having comprehensive capacity of dealing imprecise data,is applied on both coefficients for optimal fusion.By utilizing Fuzzy Rules,both Pixel and Region based fusion is executed here.Pixel Level Fusion is applied on HFS coefficient based on the energy and entropy of every pixel,according to the contourlet decomposition characteristic,so that it can be completely reconstructed.While on LFS coefficient decision map based fusion rules is applied based on the weighted average and selection max-min method,in accordance to contourlet transform.The decision is purely based on the salient factor of the LFS.Finally,like formal transform based fusion method,inversed contourlet transform is applied to get the required fused image.Four different sets of CT and MRI images of human brain are taken here for experimentation.Results obtained from various experiments shows superiority from other proposed schemes for multi-modality medical images.Both the subjective and objective parameters were analyzed and compared with different techniques which indicate that the provided schemes is efficient and have better performance of fusion.