Evaluating the Performance of Traditional Handmade Tile Workshops to Minimize Color Variation Using K-Mean Clustering and Comparative Histograms

Color variations in traditional multi-colored tiles are valued for their unique character and authenticity. Often, tiles placed side by side show noticeable color differences. This study evaluated the performance of eight traditional handmade tile workshops in minimizing these differences by comparing similar colors on adjacent tiles. Color readings were obtained using a characterized scanner, and the K-mean clustering technique was employed for indexing and comparison. The color differences between identical colors on adjacent tiles were quantified, utilizing a histogram as a judgment criterion. The scanner demonstrated reliable data uniformity with a measurement error of less than 1%. Temporal homogeneity analysis showed variations within ±1 unit, indicating negligible repeatability error. A five-unit color difference, based on industrial experience, was set as the tolerance limit. Ultimately, workshops W5, W7, W6, and W4 were evaluated further for final decision-making based on other parameters.