Fusion of Brain MR Images for Tumor Analysis using Bi-Level Stationary Wavelet Transform

Abstract

The objective of image fusion is to merge the information of two or more images to get a fused image having more distinct features for image analysis. It is very important to know the features of a brain tumor in the early stage before metastasis to save the life of a patient. Images captured at different time of same scene gives different information. Fusion of such images may be more informative for better analysis by human and even by artificial intelligent system. In recent studies the different image fusion methods have been developed both in spatial and transform domain. The fused image obtained by considering spatial domain method will have spatial distortions and due to which loss of information may happen. Such spatial distortions may be overcome by using wavelet transformation-based fusion. This paper is about the fusion of T1, T2 weighted and Flair brain magnetic resonance imaging (MRI) images using 2-level Stationary Wavelet Transformation (SWT). The image obtained after fusion is evaluated with the Entropy, Mutual Information and Fusion Symmetry and compared with the previous work done by another researcher. Experimental results show that the Entropy is improved which in turn have more information compared to the image having low entropy.