Saliency-Guided Local Full-Reference Image Quality Assessment

Research and development of image quality assessment (IQA) algorithms have been in the focus computer vision processing community for decades. The intent IQA methods is to estimate perceptual digital images correlating as high possible with human judgements. Full-reference algorithms, which full access distortion-free images, usually contain two phases: local estimation pooling. Previous works utilized visual saliency final pooling stage. In addition this, was weights weighted averaging scores, emphasizing regions that are salient observers. contrast this common practice, applied computation study, based on observation determined both by degradation simultaneously. Experimental results KADID-10k, TID2013, TID2008, CSIQ shown proposed method able improve state-of-the-art’s performance at low computational costs.