Comparative analysis of image fusion performance evaluation methods for the real-time environment monitoring system

Abstract

We discuss and compare several objective measures used for image fusion algorithm performance evaluation. Subjective assessments are given as well. Many of the considered evaluation methods originate from prior literature, we also introduce measure based on Jensen-Shannon divergence and a simple gradient-based measure, particularly well fitted for the real time fusion evaluation issue. Along with several well known fusion methods we put under tests recently developed, promising algorithm based on the fast and adaptive bidimensional empirical mode decomposition. © 2013 Springer-Verlag.