Multi-focus Image Fusion: Wavelet Decomposition and PSF Model Integration

This paper introduces multi-focus image fusion techniques, starting with wavelet transform fusion algorithms. While wavelet transforms offer non-redundancy and minimal high-frequency loss for effective fusion, they suffer from shift-variance and edge information degradation during reconstruction due to external interference. To overcome these limitations, we propose a hybrid method integrating wavelet decomposition with Point Spread Function (PSF) modeling. The approach involves non-downsampled wavelet decomposition to maintain source image dimensions, superposition of multi-directional/multi-scale high-frequency components, and feature extraction for sharp/blur target identification. The fusion algorithm design incorporates strategic source image combination while bypassing inverse wavelet transforms, yielding superior results through optimized edge preservation and reconstruction stability.