The problem of increasing efficiency of blind image quality assessment is considered. No-reference image quality metrics both independently and as components of complex image processing systems are employed in various application areas where images are the main carriers of information. Meanwhile, existing noreference metrics have a significant drawback characterized by a low adequacy to image perception by human visual system (HVS). Many well-known no-reference metrics are analyzed in our paper for several image databases. A method of combining several noreference metrics based on artificial neural networks is proposed based on multi-database verification approach. The effectiveness of the proposed approach is confirmed by extensive experiments.
This paper addresses the problem of assessing full-reference visual quality of images. A correlation between the obtained array of mean opinion scores (MOS) and the corresponding array of given metric values allows characterizing a correspondence of the considered metric to HVS. For the database TID2013 intended for a metric verification, a Spearman correlation is about 0.85 for the best existing HVS-metrics. A simple way to improve an efficiency of assessing visual quality of images is to combine several metrics: as a product of two existing metrics in certain powers that can be optimized or applying more complex structures to unify more than two visual quality metrics. We show that clustering methods can be efficiently used for this purpose. This method provides essentially larger improvement of a combined metric performance compared to the method based on their multiplication. Besides, our work specially addresses assessing images with multiple distortions. There are two such types in the modified LIVE database and two others in TID2013. Spearman rank order correlation coefficient (SROCC) between a combined metric and mean opinion score for a considered database serves as a criterion for the metric optimization. As the result of our design, the SROCC reaches 0.95 for the verification set of the database TID2013. This is considerably better than for any particular metric employed as an input where FSIMc is the best among them.