Currently, screen content images (SCIs) are widely used in our modern society. However, since SCIs have distinctly different properties compared to natural images, traditional quality assessment methods of natural images cannot precisely evaluate the quality of SCIs. Thus, we propose a blind quality evaluation method for SCIs based on regionalized structural features that are closely relevant to the intrinsic quality of SCIs. Firstly, the features of textual and pictorial regions of SCIs are extracted separately. For textual regions, since they contain noticeable structural information, we propose improved histograms of oriented gradients extracted from multi-order derivatives as structural features. For pictorial regions, since human vision is sensitive to texture information and luminance variation, we adopt texture as the structural feature; meanwhile, luminance is used as the auxiliary feature. The local derivative pattern and the shearlet local binary pattern are used to extract texture in the spatial and shearlet domains, respectively. Secondly, to derive the quality of textual and pictorial regions, two mapping functions are respectively trained from their features to subjective values. Finally, an activity weighting strategy is proposed to combine the quality of textual and pictorial regions. Experimental results show that the proposed method achieves better performance than the state-of-the-art methods.
This is an open access article distributed under the Creative Commons Attribution License
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited