Objective Quality Assessment of Screen Content Images by Structure Information

In this paper, we propose a novel full-reference objective quality assessment metric of screen content images by structure information. The input screen content image is first divided into textual and pictorial regions. The visual quality of textual regions is predicted based on perceptual structural similarity, where the gradient information is used as the feature. To estimate the visual quality of pictorial regions, we extract the luminance and structure features as feature representation. The overall quality of the screen content image is measured by fusing those of textual and pictorial parts. Experimental results show that the proposed method can obtain better performance of visual quality prediction of SCIs than other existing ones.