Open AccessStudy of local assessments of contrast for digital images
Author(s) -
Ю. И. Голуб,
Ф. В. Старовойтов
Publication year - 2019
Publication title -
sistemnyj analiz i prikladnaâ informatika
Language(s) - English
Resource type - Journals
eISSN - 2414-0481
pISSN - 2309-4923
DOI - 10.21122/2309-4923-2019-2-4-11
Subject(s) - contrast (vision) , weibull distribution , set (abstract data type) , mathematics , artificial intelligence , data set , image contrast , digital image , statistics , pattern recognition (psychology) , computer science , image (mathematics) , image processing , programming language
The article study notion of the digital image contrast. Various quantitative estimates of the contrast of digital images are investigated. Comparative experimental studies of different contrast estimates were performed on two standard image databases. It is shown that the evaluation of contrast by calculating the arithmetic average or MSE values of a set of local estimates does not always coincide with the visual contrast assessment. Experimental results showed that the Weibull distribution shape parameter is a more accurate estimate of the set of local contrast estimates calculated by the BEGH, GORD, LOEN functions and correlates well with the visual contrast estimates.