Full-reference Screen Content Image Quality Assessment by Fusing Multilevel Structure Similarity

Screen content images (SCIs) usually comprise various types with sharp edges, in which artifacts or distortions can be effectively sensed by a vanilla structure similarity measurement full-reference manner. Nonetheless, almost all of the current state-of-the-art (SOTA) metrics are “locally” formulated single-level manner, while true human visual system (HVS) follows multilevel manner; such mismatch could eventually prevent these from achieving reliable quality assessment. To ameliorate this issue, article advocates novel solution to measure “globally” perspective sparse representation. perform assessment accordance real HVS, abovementioned global metric will integrated conventional local ones resorting newly devised selective deep fusion network. validate its efficacy and effectiveness, we have compared our method 12 SOTA methods over two widely used large-scale public SCI datasets, quantitative results indicate that yields significantly higher consistency subjective scores than leading works. Both source code data also publicly available gain widespread acceptance facilitate new advancement validation.