Traditional woven fabric motifs from East Sumba represent a cultural heritage rich in aesthetic, symbolic, and philosophical values. However, the process of motif identification is still often done manually, which is time-consuming and prone to errors due to subjectivity and limited visual knowledge. This study aims to apply the Convolutional Neural Network (CNN) method to detect and classify four types of East Sumbanese fabric motifs: chicken, bird, crocodile, and horse. The dataset used consists of 200 RGB color images divided equally into training and test data. The CNN architecture used is MobileNetV2 due to its advantages in efficiency and accuracy of visual pattern recognition. To improve model performance and generalization, image augmentation techniques are used. The training process was carried out on the Google Colab platform, while model evaluation was carried out using a confusion matrix and classification reports with precision, recall, and f1-score metrics. The test results showed that the model was able to recognize motifs with varying accuracy in each class, with a total accuracy of more than 70%. These findings indicate that CNN can be an effective solution in supporting cultural preservation through the digitization and classification of traditional fabric motifs from East Sumba.
Anapaki et al. (Wed,) studied this question.