Performance Assessment of Gaussian Filter-Based Image Fusion Algorithm

Image fusion plays a vital role in many fields. Especially, fusion of infrared and visible images has high importance in every scenario from computer vision to medical sector. The objective of this work is to develop an effective method for producing clear objects with high spatial resolution along with background information by fusing infrared (IR) and visible (VIS) images. This integrated image can be efficiently utilized by humans or machines. To achieve this objective, we propose the use of Multi-Layer Bilateral Filtering (BF) and Gaussian Filtering (GF) techniques, which improvises the skewness and kurtosis of fused images. While the BF technique consistently produces higher quality images, the GF approach outperforms it by 86% in terms of statistical measures such as skewness and kurtosis. The findings demonstrate that the GF technique yields outputs with reduced noise and improved visual appeal. In this paper, we compare the assessment metrics of several outputs for both single images and a set of 100 images.