Optimized multimodal medical image fusion framework using multi-scale geometric and multi-resolution geometric analysis

For proper homoeopathic identification of the medical image, image fusion has been proposed as a mandatory solution to obtain high-spectral and high-spectral spatial data. This article presents a complete fusion system for several types of medical images according to their multi-resolution, multi-scale transforms and the Modified Central Force Optimization (MCFO) technique. Four main techniques have been proposed for this purpose; Optimized Discrete Wavelet and Dual-Tree based fusion techniques as a multi-resolution transform. Besides, the optimized Non-Sub-Sampled Contourlet and Non-Sub-Sampled Shearlet as multi-scale fusion techniques. The perfect matching between input images and minimum artifacts after image registration can be achieved through four stages in the proposed fusion algorithms. First, the input medical image is initially decomposed into their coefficients, and the MCFO method establishes the optimal gain parameter values of the resulted coefficients. Finally, the adaptive histogram equalization and the histogram matching are applied for higher clearness and better visualization of information details. The proposed algorithms are evaluated using various datasets for different medical and surveillance applications through some quality metrics. The Experimental test outcomes indicate that the proposed fusion algorithms achieve good performance with high image quality and appreciated estimation metrics principles. Moreover, it provides better image visualization and minimum processing time, which helps diagnose diseases.