A Multimodality Medical Image Fusion Method Using Region Based Approach

Abstract

Image fusion is a procedure in which the different modality images are combined to get the composite image that retains the useful information. Multimodality medical image fusion plays a vital role in diagnosis of disease and in other clinical applications. This work presents a fusion algorithm using region-based approach for medical images. The source images are partitioned using the k-means algorithm. The quadtree decomposition and Bézier interpolation techniques are utilized for reconstruction of the background. Then the background image is extracted from segmented input images to get the feature image. Feature enhancement is done to get a refined feature image. Lastly, the fused output image is found by adding the feature image with the other input image. The fusion results of the suggested scheme are compared with three other methods. The qualitative and quantitative evaluation results reveal that the proposed system performs better in comparison to the other techniques.