Comparison study of image fusion based on multi-scale transforms and sparse representation

摘要

Nowadays image fusion becomes very intensive research. Multi-scale transform and sparse representation are typical tools in fusion processing. However, most of efforts are focus on the accurate registration images. A comparative study of MST and SR method over mis-registration are proposed to investigate the robust performance. Five metrics are adopted to evaluate the quality of different methods. The experimental results suggest that the kinds of transform play a crucial role for MST methods, and SR generally shows better performance.