Влияние цветового пространства при оценке качества цифровых изображений

Abstract

Objective methods and metrics for assessing image quality are often developed for grayscale images and can't be directly applied to the color images. There are a number of color spaces with luminance component, like YUV, HSV, CIELab, HunterLab, which can be used to convert color image to grayscale, but we couldn't find any scientific papers that are devoted to the problem of what color space must be used for this conversion and how it affect the accuracy and reliability of the image quality assessment. In this paper we will try to find an answer to the question about existence of the any relationship between the particular method of calculating luminance (selecting a color space) and the adequacy of the image quality assessments.