Multi-Exposure Fusion
Multi-Focus Image Fusion Based on Mallat Algorithm
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.
Performance Evaluation of Different Fusion Methods for Multi-focus Image Fusion
To objectively and quantitatively evaluate the performance of various fusion methods in multi-focus image fusion, this study analyzes images based on their statistical characteristics. Without reference standard images, four key parameters are selected for comprehensive assessment: average gradient (sharpness), spatial frequency, information entropy, and standard deviation, which collectively measure fusion method performance and can be implemented through computational algorithms.