An Efficient Quality Assessment Method for Screen Content Image Based on Gabor

The goal of an objective quality assessment metric is to use a computational model to measure the image visual quality, which consistent with the subjective judgment of the human being. Accurately evaluating the visual quality of the image is an important task under many application scenarios, such as image compression, image reconstruction, and image enhancement. In this paper, we present an accurate full-reference image quality assessment (IQA) metric for screen content (SC) images. Compared to the other classical algorithms, the proposed method is based on convolution operations, which extremely reduced computational complexity. First, the Gabor filter is used to capture the image structure information, usually composed of edges and contours, which is used to understand the image. And then we detect the visual sensitive areas that can attract more attention from the human visual system (HVS). Finally, the map of visual sensitive areas is used for the weighting strategy. Extended experiments on the SIQAD database validate the superiority of the proposed method. Specifically, it is more consistent with subjective evaluation results than the most popular and state-of-the-art IQA metrics and requires lower computational complexity
¹
.