Batik Pattern Classification Using Decision Tree Based on Color-Texture Features

Indonesia exhibits a wide array of conventionally crafted textiles characterized by their regional distinctiveness. East Kalimantan, a province in the Republic of Indonesia, is renowned for its traditional textile craft called Batik. The Batik produced in East Kalimantan often incorporates a diverse spectrum of hues, such as orange, green, pink, and red. The cultural themes of the Dayak people are mostly shaped by their viewpoints and philosophies of nature and the surrounding environment. Regrettably, a considerable segment of the indigenous community remains uninformed about East Kalimantan Batik, hence impeding their ability to recognize and discern its distinct characteristics. Hence, the utilization of an automated image processing technique became necessary in order to detect and categorize the predominant themes found in East Kalimantan Batik. In order to optimize classification outcomes using this strategy, it is vital to utilize appropriate and discerning features. The objective of this study is to examine the characteristics of color and texture to generate separate attributes. The color characteristics employed the occurrences of applied color inside the RGB color spaces. In order to extract textural information, the Gray Level Co-occurrence Matrix (GLCM) was utilized. The collected features were subsequently included in the Decision Tree classifier. The dataset employed in this analysis comprises 400 batik images, encompassing 100 images each of Batang Garing, Burung Enggang, Shaho, and Tameng. Through the utilization of color features, the methodology achieved a significantly high level of performance, attaining an accuracy rate of 99.9%.