<![CDATA[Batik is an Indonesian cultural heritage known for its diverse motifs; however, manual classification of these motifs remains a significant challenge. This study aims to develop a batik motif classification model using the ResNet50 architecture enhanced with data augmentation to improve model accuracy. The dataset consists of four batik motif classes: Kawung, Mega Mendung, Parang, and Truntum. In this research, the model was trained using fine-tuning on ResNet50, with additional CNN layers for feature extraction. The results demonstrated that the proposed model achieved a highest accuracy of 97.80% on test data and 96.80% on validation data, significantly outperforming methods without data augmentation. Researchers will also compare accuracy with other deep learning models for classifying Keraton Batik images. This study concludes that applying a fine-tuned ResNet50 model with additional CNN layers and data augmentation effectively classifies batik motifs, offering substantial potential to automate batik motif recognition and supports digital preservation and development of batik in the cultural and creative industries.]]>
Copyrights © 2026
