<![CDATA[Batik is one of Indonesia's cultural riches, characterized by diverse patterns and deep philosophical values rooted in each region. However, the manual process of identifying and classifying batik motifs still requires specific expertise and is often time-consuming. This study aims to implement the ResNet50 architecture in designing a model for recognizing Indonesian batik motifs using a Convolutional Neural Network (CNN) approach. The model is trained to differentiate four batik motif categories: Batik Corak Insang, Batik Dayak, Batik Ikat Celup, and Batik Megamendung. Experimental results show that the model achieves an accuracy rate of 81%, with a validation accuracy of 80,68% and a validation loss of 66,94%. Among all classes, the model performs best in classifying Corak Insang and Dayak motifs. The trained model is deployed in a web-based application that enables users to upload batik images and receive instant classification results. Based on these outcomes, it can be concluded that the ResNet50 architecture, when integrated with CNN, can be effectively utilized to support automatic batik motif recognition and contribute to the digitalization and preservation of local cultural heritage through artificial intelligence technology]]>
Copyrights © 2025
