Robust Multi-Focus Image Fusion: A Novel Framework with Optimized Performance and Efficiency

Abstract

This paper presents a new image fusion technique that employs shearlet-based edge enhancement and adaptive fusion rule to improve the visual quality and information retention of fused images. The new method first applies shearlet-like edge detection filters to the input images with emphasis on finer details such as edges. The enhanced images are then adaptively fused based on local image properties, including texture and edge sharpness, to obtain the fusion weights. The performance of the method is rigorously tested using quality measures such as Mutual Information (MI), Entropy, Peak Signal-to-Noise Ratio (PSNR), and Structural Similarity Index (SSIM). The execution time of each algorithmic process (enhancement, fusion, and computation of metrics) is also recorded, which provides information on its efficiency. The experimental results confirm that the new method achieves high-quality fusion with preservation of edges and finer details and retention of essential information from the original images. The method is promising for applications in medical imaging, remote sensing, and computer vision, where image clarity and preservation of details are significant.