Designing a Web Application for Detecting Indonesian Batik Motifs Based on Image Processing and Machine Learning

Abstract

Machine learning batik motif detection is important because it helps identify, classify, and find batik by motif and area. The diversity of batik motifs in Indonesia poses a challenge to society, as many motifs have similarities in pattern or color, leading to errors in identification. Researchers have used machine learning techniques to address this problem. Machine learning models with image preocessing techniques such as torch techniques, log gabor, gray level co-occurrence matrix (GLCM) techniques have been used to identify batik motifs with high accuracy. This application will be developed using the web information system development methodology (WISDM) methodology. These advances in machine learning of batik motif detection contribute to preserving Indonesian culture and heritage. The best results were obtained from the combination of gabor, log gabor, GLCM features with retrieval rate quality reaching 84.54% in motif detection.