Perceptually Correlated Image Quality Metrics
Jonathan B. Phillips, Henrik Eliasson · 2017
The aim of Chapter 7 is to provide a link between the objective metrics presented in Chapter 6 and results from subjective studies as presented in Chapter 5. In order to do this, a simple model of the human visual system needs to be considered. Therefore, a part of the chapter explains relevant parts of the human visual system. A few of the most important local attributes, sharpness, texture blur and noise, are discussed in the context of defining a framework for obtaining perceptually correlated metrics for these attributes. Other important image quality metrics developed from human visual models, such as S-CIELAB and the iCAM framework, are presented. Color and the difficulties of obtaining a good color reproduction benchmark are also discussed.