Multimodal Medical Image fusion with Butterworth filter in NSCT Domain based on dual fusion frame work

Abstract

Multi-modal fusion of medical images gives more sophisticated and comprehensive information in the latest medical diagnosis. Image fusion has attracted the attention of researchers because, significant advances in digital image applications in key fields like medical, space research, remote sensing. The aim of fusion is to get complementary data of different images. This article deals with the fusion of medical images at two levels by utilizing the features of DWT and NSCT independently at first and spatial approach at next level using different fusion rules at both levels that enhance the features of fused image for addressing various lesions of organs. In this work, two medical images from two modalities are taken. After preprocessing and filtering, they will be integrated according to DWT and NSCT transformations using different fusion rules to get new images. These new images are again fused in spatial domain to produce high featured image. Phase congruency and directive contrast of decomposed sub-bands in NSCT are taken to pick useful bands. Absolute maximum values of decomposed sub-bands of DWT are considered to elect useful bands of DWT. Outcomes of DWT and NSCT are taken as source images for next level fusion. These images are fused in spatial domain using edge based similarity measure rule. In this way the more useful information is presented. The results of this work are compared in terms of quality of mutual information, entropy, and edge based similarity measure qualitatively and quantitatively. The performance parameters give tremendous improvements