Pixel saliency metric based pixel-level medical image fusion

This paper proposes a novel medical image fusion algorithm based on pixel salience metric in gradient domain. We first obtain the gradient image of every input image and measure their salience map, and then use saliency to weight gradient pixel to acquire a target gradient field. Applying dynamic range compression (DRC) algorithm to reduce halo effect. Finally, exploiting successive over-relaxation method to solve a Poisson equation for minimizing the energy function. Experimental result shows that the proposed method is the best in the concept of saliency compared to the existing pixel-level image fusion methods.