Swarm intelligence based image fusion for noisy images using consecutive pixel intensity

Abstract

A novel image fusion technique is presented, aiming at resolving the fusion problem of noisy images. In this paper, a new activity level measurement based on consecutive pixel intensity similarity is proposed for detecting the noise free and noisy parts from the source images and also the fusion technique is optimized using particle swarm optimization for obtaining the optimized fused image. Experiments have been made on images affected by Gaussian noise, salt and pepper impulsive noise, speckle noise and Poisson noises for examining the efficiency of the proposed algorithm. The proposed framework is evaluated using quantitative metrics such as root mean square error, peak signal to noise ratio, mean absolute error, percentage fit error, structural similarity index and mutual information. The experimental results demonstrate the outperformance of the proposed algorithm over many other well known state-of-the-art fusion techniques reported in the literature.