ANALYSIS OF IMAGE BRIGHTNESS BASED ON BOX-COX TRANSFORMATION AND VISUAL EVALUATION

Abstract

This article discusses methods for analyzing the brightness characteristics of digital images using statistical and visual methods. Image files are first converted to grayscale format and normalized using Box-Cox transformation. Then, metrics such as average brightness, dispersion, skewness, kurtosis, entropy, sharpness, and contrast are calculated. Also, a simulated MOS (Mean Opinion Score) score is generated for each image. The analysis results are visualized using scatter plot, boxplot, histplot, pairplot, and heatmap. Spearman and Kendall correlations are used to determine the relationship between metrics and MOS. This approach is recommended as an effective tool for assessing digital image quality and statistical analysis.