ABSTRACT
With 300 figures, tables, and equations, this book presents a unified approach to image quality research and modeling. The author discusses the results of different, calibrated psychometric experiments can be rigorously integrated to construct predictive software using Monte Carlo simulations and provides numerous examples of viable field applications for product design and verification of modeling predictions. He covers perceptual measurements for the assessment of individual quality attributes and overall quality, explores variation in scene susceptibility, observer sensitivity, and preference, and includes methods of analysis for testing and refining metrics based on psychometric data.
Characterization and quality: can image quality be usefully quantified ; the probablistic nature of perception; just noticable differences; quantifying preference; properties of ideal interval and ratio scales; establishing image quality standards;calibrated psychometrics using quality rulers; practical implementation of quality rulers; a general equation to fit quality loss functions; scene and observer variability; predicting overall quality from image attributes. Design of objective metrics:overview of objective metric properties; testing objective metrics using psychometric data; a detailed example of objective metric design.
