Implementation and comparative analysis of various Pyramid-based Image Fusion techniques for Multimodal MRI images of brain

Image fusion has become more important in recent years in image processing because there are so many acquisition methods. Image fusion is a way to take the important features from multiple images and combine them into one (fused) image that may be more useful than the original. Image fusion is often used to integrate complementary multi-temporal, multi-view, and multi-sensor data into a single image, enhancing image quality while preserving the quality of essential features. The current study aims to apply different pyramid-based image fusion algorithms to the well-known BraTS multimodal MRI dataset. The objective of this work is to conduct a comparative study and present the analysis of these algorithms and evaluate their performance with different quality metrics such as spatial frequency, entropy, and mutual information. The findings of this work show that the fusion of Flair and T2 slices of brain MR images leads to improved quality metrics in terms of Spatial frequency, Mutual Information, and edge preservation parameters which indicates that the fused image contains more information when compared to that of other fusion combination of multi-modal MRI. Image fusion plays a critical role in medical image analysis by combining several features of the image into single representative image.