A Multi-Modal Medical Image Fusion Technique Using Structural Similarity Index Measure and Correlation Coefficient

This research paper introduces an innovative approach (M3IF-SSIM-CC) to multimodal medical image fusion (M3IF) using the Non-Subsampled Contourlet Transform (NSCT), specifically targeting enhanced clinical decision-making through improved diagnostic imagery. M3IF-SSIM-CC integrates crucial diagnostic information from Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) scans. Initially, NSCT decomposition is applied independently to CT and MRI images, generating low-pass (LP) and high-pass (HP) sub-bands. The LP sub-bands undergo fusion through a correlation-based rule designed to maintain structural coherence and maximum information retention. Concurrently, the HP sub-bands are combined using a similarity-based fusion technique aimed at preserving critical edges and details. Following the fusion of these sub-bands, inverse NSCT reconstructs a comprehensive image that encapsulates anatomical and pathological features from both modalities. Rigorous evaluation, conducted using various quantitative metrics and clinical assessments, demonstrates the superiority of the M3IF-SSIM-CC over traditional fusion techniques. Results show significant improvements in image clarity, preservation of detail, and overall diagnostic accuracy, facilitating more precise and confident clinical interpretations. M3IF-SSIM-CC represents a meaningful advancement in medical imaging technologies, providing clinicians with enhanced diagnostic tools for improved patient outcomes.