Image Fusion Using Optimization of Statistical Measurements

The purpose of image fusion is to create a perceptually enhanced image from a set of multi-focus or multi-sensors images. In the methods we are about to describe we do not a priori know the ground truth image: these are blind fusion methods. There are mainly two groups of fusion methods depending on the signal domain they are applied: spatial domain methods and transform domain methods. The Dispersion Minimisation Fusion (DMF) and Kurtosis Maximisation Fusion (KMF) based techniques we are going to discuss are spatial domain methods that is to say the fusion is simply performed on the image itself. In this work we propose to linearly combine the input images with appropriate weights estimated using specific mathematical performance criteria which evaluate in various ways improvement in visual perception. More specifically, in order to estimate the weights we propose iterative methods which use cost functions based on two statistical parameters, i.e., the dispersion and the kurtosis. The optimisation of the proposed cost functions enables us to obtain a fused image which is less distorted compared to the input ones.