A Comprehensive Review of Medical Image Fusion Algorithms

Abstract

The challenge of manual design can be overcome by deep learning models, which can automatically extract the most useful elements from data. Introducing a deep learning model to the picture fusion field is the aim of this paper. Using supervised deep learning, it aims to create a novel concept for picture fusion. Pattern recognition and image processing are two fields where deep learning technology has been thoroughly investigated. The characteristics of multi- modal medical images, medical diagnostic technology, and practical implementation will be taken into consideration when proposing a multi-mode medical image fusion with deep learning, in accordance with the practical requirements for medical diagnosis. It is not only capable of compensating for the shortcomings of MRI, CT, and SPECT image fusion; it can also be applied in batch processing mode to various multi-modal medical image fusion problems and successfully get over the restriction of one-page processing. By obtaining the salient characteristics in the fusion results, medical image fusion techniques are used to improve the quality of medical images. As a result, they improve medical imaging' clinical usability for diagnosis and appraisal issues. In order to accomplish this, the fusion result must capture the complimentary information found in two or more images with various modalities. Aim of this paper is to provide readers with the information of the process and steps taken in medical image fusion and medical image classification of various diseases.