A Multi-scale Structure SIMilarity metric for image fusion qulity assessment

Based on the fact that the human visual system is not only highly adapted for exacting structural features such as lines, edges, contours from the input images, but also has characteristics of multi-channel (multi-scale) information processing, a Multi-scale Structure SIMilarity(MSSIM) metric to assess image fusion algorithms is proposed. Compared with the existent single-scale assessment metrics, the proposed metric provides more flexibility on account of considering the variations of viewing conditions and has better consistence with human perceptions. Visual experiments and quantitative analysis confirm its effectiveness. Some medical image fusion examples demonstrate its promising applications.