<![CDATA[Batik is an Indonesian cultural heritage recognized by UNESCO as a Masterpiece of the Oral and Intangible Heritage of Humanity. One distinctive regional variation is Batik Lontara from South Sulawesi, which features motifs derived from the traditional Bugis–Makassar Lontara script. However, the visual similarity between woven and non-woven (stamped) Batik Lontara makes manual identification difficult for the general public. This study proposes an image-based classification approach to distinguish woven and non-woven Batik Lontara using texture analysis and machine learning techniques. Texture features were extracted from batik images using the Gray Level Co-occurrence Matrix (GLCM), including contrast, homogeneity, energy, and entropy. The extracted features were then classified using the k-Nearest Neighbor (k-NN) algorithm with varying values of k (5, 7, and 9). The dataset consisted of Batik Lontara images captured using a smartphone camera and divided into training and testing sets using both hold-out and k-fold cross-validation schemes. Experimental results show that the highest classification accuracy of 86% was achieved using the k-NN algorithm with k = 9 under 5-fold cross-validation, while the hold-out method produced the best accuracy of 87% at k = 5. These results demonstrate that the proposed GLCM and k-NN-based approach is effective for classifying Batik Lontara types and has strong potential for supporting cultural heritage preservation through digital image processing.]]>
Copyrights © 2026
