Synthetically evaluation system for multi-source image fusion and experimental analysis (Q879700)

The authors provide a brief account of three sets of performance assessment indexes for image fusion: independent single evaluation indexes, union single evaluation indexes and synthetic evaluation indexes. They also inform that they have implemented several of them in an image evaluation module for the ENVI/IDL platform. Three application examples of wavelet fusion of several types of images are analyzed with this tool. The authors come to the conclusion that the union single evaluation indexes are the ones that best account for the observed changes in the fusion of images coming from different sources, but they acknowkedge that the synthetic evaluation indexes are a new and promising field of research and application, since they provide a framework for combining subjective and objective judgement in a quantitative manner. The authors unnecessarily define well known statistics as, for instance, sample mean and variance, while other not so well known quantities are used without a proper presentation, discussion or references to accessible literature. The synthetic evaluation indexes are only described without definitions or adequate references. The paper is, nevertheless, interesting for image data users having access to information from different sensors over the same area.