Assessment of digital image quality for visual objects in Gamified Systems

Abstract

This paper devoted to the development and justification of a comprehensive approach to assessing the quality of digital images in gamification environments, taking into account the peculiarities of the perception of visual objects by users. The relevance of the study is due to the growing role of gamified visual interfaces in digital applications and the need for objective quality control of visual content, which directly affects the level of user engagement and motivation. The paper analyzes modern methods for assessing the informativeness and quality of digital images, in particular, approaches based on texture analysis, spatial structure, machine learning, and deep neural networks. A hybrid method is proposed that combines the evaluation of color differences according to the CIEDE2000 standard with fractal analysis of structural characteristics of images, which allows taking into account both color and geometric-textural properties of visual objects. The research methodology is based on the formulation and testing of hypotheses regarding the influence of image processing quality on the perception of objects in gamified environments, mathematical modeling of the integral quality indicator using weighting factors, as well as software implementation of the proposed approach using Python. Practical testing was carried out on a set of digital images of different levels of graphic complexity (simple, medium complexity, and complex), for which the CIEDE2000 indicators, fractal dimension, integral quality index, and MSE and SSIM metrics were calculated. The results obtained confirm that the combination of color and structural analysis provides a more complete and objective assessment of the quality of gamified images compared to the use of separate methods. The work shows that improving the quality of images is statistically associated with improving the perception of visual objects, and the use of a combined method allows for more correct differentiation of quality levels for images of different complexity. The proposed approach can be used for automated quality control of visual content in computer games, educational platforms and other gamified systems, and also creates a basis for further research in the direction of adaptive and intelligent assessment of visual data.