A novel approach for medical image fusion using fuzzy logic type-2

Medical image fusion is used to integrate the essential features present in different medical images into a single image to improve the clinical accuracy to take better decisions. Multimodal medical image fusion combines the images obtained from different modalities like Positron Emission Tomography (PET), Computed Tomography (CT), Magnetic Resonance Imaging (MRI) and others. CT scan provides detailed information on bony structures whereas MRI scan provides details on soft tissues. Fusion of these images is useful for doctors to diagnose and plan treatment for patients. Various methods have been proposed during recent years for medical image fusion but these methods suffer from the issue of fusion performance due to different sensor modalities. In this paper, we validate that for such images, fusion performance can be improved using a learning based fusion scheme which uses training and testing phase by considering pixel wise processing of image. In order to improve image fusion performance, here we implement fuzzy logic type-2 based approaches for medical image fusion for CT and MRI images. Final fused image is achieved by applying rule based method which includes fuzzification inference, type reduction, and defuzzification of the input image. Experiments are conducted by applying fuzzy logic type-1, neural network, Neuro-fuzzy approach and fuzzy logic type-2 based fusion scheme. Performance of image fusion based on Fuzzy Logic Type-2 is compared with other state-of-art techniques using various performance metrics such as Mutual information, standard deviation and edge based similarity. Experimental results shows that fuzzy logic Type-2 performs better when compared to neural network based approach.