A compendious analysis of feature-extraction algorithms to frame fusion rules

Abstract

Advancement in sensor technology provides the complete information captured by multiple sensors. To reduce the eye strain and workload from analyzing the scene with multiple images, the information is combined into a single image by the process called image fusion. In this paper, a compendious analysis of feature-extraction based fusion algorithms that define an appropriate fusion rule is reviewed. A state-of-art classification of feature-based fusion schemes is carried out and the extracted feature maps are presented. The qualitative analysis for different fusion methods are illustrated and compared. The quantitative fusion metrics are grouped as contrast, information, edge and visual based metrics and are evaluated. Finally, the conclusion and future directions are briefed out.