A New Approach to Assessing the Perceived Contrast

Abstract

This paper is devoted to the problem of increasing the efficiency of no-reference assessment of the perceived contrast of images. The estimate of perceived contrast of the image is based on measuring the contrast of the image elements (objects) relative to each other. Contrast definitions for a pair of objects in an image are the basis for developing the no-reference metrics of image integral contrast. This paper proposes a new approach to quantify the perceived contrast for two objects in an image by assessing the contrast of these objects in a linearly-transformed (used as reference) image with a specified range of brightness values and subsequent normalizations of the obtained result of measurement. The range of brightness for the reference image are chosen in such a way as to provide an accurate measurement of the perceived contrast with using its definitions based on the Weber-Fechner law. Based on this approach a new generalized description of the perceived contrast of two objects in the image is presented. Proceeding from this description, new definitions of the weighted and relative contrast of a pair of objects in the image are proposed. Research results confirm the effectiveness of the proposed approach to quantify the perceived contrast of images.