Classification of North Sumatra Batak Ulos Based on Ethnicity Using Convolutional Neural Network Algorithm Approach

Abstract

Ulos is a piece of cloth produced through a weaving process that reflects a rich cultural heritage and has high value. The patterns contained in woven ulos often contain philosophical meanings, reflecting the traditional values, beliefs and history of the communities that produce them. However, in reality there are still many Batak young men and women and the general public who are not yet able to distinguish between types of ulos. This research aims to help identify types of uos with the hope of providing deeper insight into the diversity of ulos based on ethnicity in North Sumatra. The dataset used in this research consists of 600 datasets which are divided into 6 types of ulos. Before the classification process is carried out, the data is cleaned through data preprocessing by cropping the image data to produce the same image data size. The research results show a classification accuracy rate of 96%. This finding confirms that the Convolutional Neural Network (CNN) method can be applied to classify ulos based on ethnicity. This has important implications in increasing understanding and appreciation of the traditional arts of the Batak tribe and supporting efforts to preserve this valuable cultural heritage