Classification of Batik Image using Grey Level Co-occurrence Matrix Feature Extraction and Correlation Based Feature Selection

Abstract

Batik is a cultural heritage that has become part of Indonesian society. Batik has a variety of patterns and motifs. Each region has varieties of motifs in terms of color, texture and production techniques. This study discusses the feature selection method for classification of batik image into Kawung, Lereng, Nitik and Tambal. Selection of the right features by eliminating redundant features can result in higher accuracy. Another important step is feature extraction. This research applies the Gray Level Co-occurrence Matrix feature extraction to extract features in the image of batik. The total features obtained by extracting batik images using GLCM are 20 features. From 20 features, CFS is able to reduce 70% of irrelevant features. The results showed that the classification of batik using Backpropagation resulted in an accuracy of 83% and the classification using the K-Nearest Neighbor method was 67%.