PENERAPAN DEEP LEARNING UNTUK PENGENALAN CITRA TULISAN TANGAN AKSARA LAMPUNG

Abstract

In the field of artificial intelligence, technology has opened new opportunities for preserving cultural heritage, including the Lampung script. One method that shows great potential in script recognition is Deep Learning. This research aims to develop a handwritten image recognition model for the Lampung script using deep learning methods with the ResNet- 50 architecture and transfer learning. Handwritten data of the Lampung script were collected from various sources and processed through image normalization and augmentation to improve the quality of the training data. The model was built using the ResNet-50 architecture with transfer learning, which allows the use of pre-trained models to enhance efficiency. The research results show that the model is capable of classifying 20 primary characters of the Lampung script with adequate performance. However, the study suggests further data collection and exploration of augmentation techniques to improve results. This research not only aims to enhance the recognition of the Lampung script but also contributes to efforts to preserve and introduce the Lampung script to the wider public through modern technology.