Modeling image quality attributes based on image parameters for mobile displays

Abstract

High image quality (IQ) on the widely used mobile displays is extremely desirable for viewing static and moving images. This study focuses on modeling the perceived IQ considering all types of images for mobile displays. Large-scale psychophysical experiments via categorical judgment method were carried out on mobile displays of different physical sizes under lighting conditions of two illuminance levels, in order to visually assess seven perceptual attributes, i.e. naturalness, colorfulness, brightness, contrast, sharpness, clearness, and overall IQ. The results reveal that clearness and naturalness are two principal attributes to predict overall IQ for natural scene images, whereas clearness and colorfulness are key attributes for other types of images (games, maps, and internet). Moreover, the influences of display technology, display size and lighting level have been analyzed as well. Based on the preliminary models of overall IQ with these selected attributes, further expressions to predict individual IQ attributes could be modeled by combining some image and display parameters.