Premium
85‐4: A Subpixel‐based Objective Image Quality Metric with Application to Visually Lossless Image Compression Evaluation
Author(s) -
Cook Gregory W.,
Ribera Javier,
Stolitzka Dale,
Xiong Wei
Publication year - 2018
Publication title -
sid symposium digest of technical papers
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.351
H-Index - 44
eISSN - 2168-0159
pISSN - 0097-966X
DOI - 10.1002/sdtp.12103
Subject(s) - subpixel rendering , lossless compression , artificial intelligence , computer vision , metric (unit) , computer science , standard test image , human visual system model , image quality , observer (physics) , image compression , mathematics , pattern recognition (psychology) , data compression , image (mathematics) , image processing , pixel , operations management , physics , quantum mechanics , economics
Subjective Test Scoring is the “gold standard” for determining whether two images are considered visually lossless, and S‐CIELAB is a good tool for guiding researchers in finding impaired regions. In this paper we introduce a new metric, which we call ISETBIOLAB, that is based on the computational observer model and is similar in spirit to S‐CIELAB but also incorporating the human visual system model. It is shown to have a better correlation to Subjective Test Scoring for visually lossless compression evaluation than either PSNR or S‐CIELAB. Also, since it is based on spatial and spectral expansion of subpixels, it is useful for evaluating any subpixel arrangement, and thus allowing direct comparisons of different display types.