Dual-Anchor Metric Learning for Blind Image Quality Assessment of Screen Content Images

The artificial portion of screen content images (SCIs) destroys the natural scene statistics of natural scenes. It is also impractical to obtain an accurate statistical model due to the variable composition of the artificial and natural parts in SCIs. To resolve this problem, this paper presents a dual-anchor metric learning (DAML) method that is inspired by metric learning to obtain discriminative statistical features, and further to identify complex distortions, and predict SCI image quality. First, two available statistical models with prior data are constructed as the target anchors of the statistical model from two uncorrelated pristine databases, which can effectively enhance the metrical discrimination of the mapping relation between the feature representation and quality degradation by conditional probability analysis. Then, the distances of high-order statistics are softly aggregated to conduct metric learning between local features and clusters of each target statistical model. Through empirical analysis and experimental verification, only variance differences are used as quality-aware features to benefit the balance of complexity and effectiveness. Finally, mapping model between the target distances and subjective quality can be obtained by support vector regression. To validate the performance of DAML, multiple experiments are carried out on three public databases, and the results confirm the superior performance of the proposed method compared to existing relevant blind image quality assessment methods with SCIs