<![CDATA[Cartoons are characterized by simple, flat, and solid colors, making them appealing for various applications, such as avatars or entertainment. However, the manual cartoonization process demands high artistic skill and is time-consuming. This research aims to automate the process through color quantization using K-Means Clustering as a solution to simplify the color palette of natural images. The main issue addressed is the selection of the optimal color mode and features to achieve the desired cartoon effect. In the methodology, the HSI (Hue, Saturation, Intensity) color mode is utilized, where K-Means clustering is specifically applied to the Hue feature only to separate color grouping from the influence of light gradation. The resulting clusters are then combined with discretized Intensity values to sharply distinguish between dark and light colors. Experimental results indicate that the K-Means algorithm is effective for color quantization, producing simpler and more solid colors that visually approximate the original color tones. This study proves that using the Hue feature in K-Means is a suitable strategy for realizing the flat color palette characteristic of cartoons. ]]>
Copyrights © 2025
