A simple and efficient hybird image fusion approach

In this paper, we propose a new pixel-based image fusion approach where fusion decision is made based on image diffusion operations. The fusion process begins with blurring the input images using linear filtering techniques to generate a smooth version of the image. Then a temple image is created based on the difference between the source images and its diffused version obtained. Consider that for points located on the homogeneity regions of the source image their corresponding points on the temple image would have little variations. However, for the points of edges in source image, their corresponding points will have large varying in intensities. Based on this, the fusion rule can be made by measuring strength of the edge from the temple image. It results in a more efficient fusion approach. The performance assessment of the proposed method is established by using mutual information measurement such as Root Mean Square Error and Peak Signal to Noise Ratio. The numerical experiment results show that the proposed method is an efficient approach to image fusion.