MIF-SWT-AD: Medical Image Fusion Using Stationary Wavelet Transform, Adaptive Feature Selection and Anisotropic Diffusion

Image fusion is a crucial technique in multi-modal imaging applications such as medical diagnostics, where integrating complementary information from different imaging modalities (e.g., Magnetic Resonance Imaging (MRI), Computed Tomography (CT), Single Photon Emission Computed Tomography (SPECT), and Positron Emission Tomography (PET) etc. to improve clinical diagnosis and decision-making. This paper presents a hybrid image fusion method (MIF-SWT-AD) based on Stationary Wavelet Transform (SWT), incorporating Variance-Based Approximation coefficients (ACs) Fusion, Local Energy-Based Detail coefficients (DCs) Fusion, and Anisotropic Diffusion (AD) Filtering for post-processing step. The MIF-SWT-AD decomposes input images using SWT to obtain multi-resolution sub-bands (ACs, and DCs). The ACs are fused using a variance-based weighting strategy, ensuring dominant structural information retention, while the detail coefficients (DCs) are fused using local energy-based selection, preserving crucial edge and texture features. Finally, AD filtering is applied to the fused image for noise reduction and edge enhancement. The experimental results demonstrate that the proposed method achieves superior fusion quality compared to conventional fusion techniques, The proposed method effectively balances feature retention, noise suppression, and visual enhancement, making it a promising approach for high-quality image fusion applications.