Comparative Analysis of Machine Learning and Contrast Enhancement Algorithms for Classifying Besurek Batik Motifs

Abstract

Batik Besurek is one of the cultural heritages of Bengkulu Province that has unique motif characteristics, requiring support from digital technology for its preservation and recognition. This study aims to compare the performance of six machine learning algorithms, namely Support Vector Machine (SVM), K-Nearest Neighbor (KNN), Decision Tree (DT), Random Forest (RF), Naïve Bayes (NB), and Logistic Regression (LR), in classifying five Batik Besurek motifs. Furthermore, this study analyzes the effect of applying contrast enhancement methods, including Histogram Equalization (HE), Contrast Limited Adaptive Histogram Equalization (CLAHE), and Gamma Correction (GC), on improving the performance of the best-performing algorithm. The dataset consists of 500 Batik Besurek images representing five motif classes, namely Kaligrafi, Rafflesia, Burung Kuau, Relung Paku, and Rembulan. All images undergo preprocessing, are transformed into one-dimensional vectors (flatten), and are divided using the hold-out validation method with an 80% training data and 20% testing data ratio. The experimental results show that SVM achieves the best performance compared to other algorithms, with a training accuracy of 86.75% and a testing accuracy of 81.00%. The application of CLAHE on SVM improves the training accuracy to 87.50% and testing accuracy to 82.00%.