Predicting the visual gloss of textured samples using the grey level co-occurrence matrix method

Gloss along with color, texture, and translucency is the main attributes of visual appearance. Conventional gloss metrics often fail to predict human perception of gloss, especially when surfaces are textured or colored. The present study investigates Gray Level Co-occurrence Matrix (GLCM) features as descriptors for the perceived gloss of colored sandpaper samples that varied in roughness and color. Visual assessments provided perceived gloss values, which were found to correlate poorly with traditional gloss measurements but were significantly influenced by both texture and color. GLCM features were extracted from sample images and correlated with perceived gloss. Strong correlations were found, notably with Information Measures of Correlation ۱ (positive, p≈ ۰.۸۰) and ۲ (negative, p≈ -۰.۷۵), Homogeneity (negative, p≈ -۰.۷۰), and Contrast (positive, p≈ ۰.۶۶). It can be concluded that GLCM features capture textural nuances associated with perceived gloss well. So image processing techniques can provide a robust perceptually relevant alternative for conventional gloss measurement methods.

کلیدواژه ها:

Gloss perception ، Texture Analysis ، Image Processing ، Surface Reflection ، Co-occurrence matrix