Investigation of Shift Dependency Effects on Multiresolution-Based Image Fusion Performance

摘要

Image fusion techniques using multiresolution decomposition have become a hot research area due to the fact that it is suitable for multi-scale properties of the human vision system. According to subsampling methods, multiresolution decompositions can be divided into redundant and non-redundant which corresponds to different shift dependency such as shift-variant or shift-invariant. However, there have been comparatively few studies that have focused on its effects on the fusion performance. This paper investigates shift dependency of various multiresolution-based fusion schemes and analyzes its effects on image fusion performance by quantitative and qualitative methods. We conduct experiments by combining 8 popular multiresolution decomposition methods such as pyramid, wavelet and beyond wavelet with two popular fusion rules. By analyzing and comparing the experimental results, we propose some guidance for using multiresolution-based fusion schemes.