A novel multi-focus image fusion algorithm using edge information and K-mean segmentation

Abstract

In this paper we introduce a new algorithm for multi-focus image fusion based on edge information of the source images and K-mean segmentation. The basic idea is to extract the edge information of the source images, divide the images into blocks and then select the blocks with higher edge information to construct the resultant fused image. The pixels in the unallocated blocks of the fused image are then selected from the source images by recursive K-mean segmentation and comparison of the corresponding segments of the source images. The major achievement of this algorithm is its excellent sharpness, clarity, superior edge preservation with minimum fusion artifacts.