Objective Evaluation of Several Shift Invariant Image Fusion Methods

Abstract

The results from most of image fusion methods are evaluated using single criteria, which has its limitations. A quantitative method is proposed to evaluate several shift invariants by using three fusion approaches, mainly the discrete wavelet frames, steerable pyramid, and dual-tree complex wavelet transform, respectively. The performances and features of these approaches are compared each other. The experimental simulation proves that the proposed evaluation is an efficient and applicable model.