Enhanced Visual Integration of Medical Images based on Gaussian of Differences

Analyzing images captured from the same source using different cameras or sensors, which may have different structures or spectral bands, is a highly intricate task. In order to resolve this problem, a single image that includes all the details from each source is necessary. This can be achieved by merging multiple images together, a technique known as image fusion. This article presents a straightforward and efficient technique for blending images based on pixel weights. The method involves utilizing edge information derived from the each pixel in the source images, determining the distance to its neighboring pixels, and using the Gaussian filter. The present method, Differences of Gaussian or Gaussian of differences (GoD), is tested via various infrared and medical images. Different techniques like existing multi-focus images, multi-exposure images, and fusion were assessed in order to compare their effectiveness in terms of objective fusion quality measures. The specific details of the parameters used for the method are also provided. An improvement has been made to the method by incorporating a pattern search (PS) algorithm, leading to a more effective optimization approach.