Handwritten Batak Toba Script Recognition Based on Deep Learning Using the Convolutional Neural Network (CNN) Algorithm

Wahyu Ardiantito Samosir; Zulfahmi Indra; Insan Taufik; Susiana

doi:10.59934/jaiea.v5i1.1795

Authors

Wahyu Ardiantito Samosir Universitas Negeri Medan
Zulfahmi Indra Universitas Negeri Medan
Insan Taufik Universitas Negeri Medan
Susiana Universitas Negeri Medan

DOI:

https://doi.org/10.59934/jaiea.v5i1.1795

Keywords:

Batak Toba script, handwriting recognition, Convolutional Neural Network, transliteration, deep learning, PWA.

Abstract

The Batak Toba script is one of Indonesia’s cultural heritages that has become increasingly rare and less recognized among younger generations. This research aims to develop a handwriting recognition system for Batak Toba characters using the Convolutional Neural Network (CNN) method, capable of accurately recognizing characters, transliterating them into Latin script, and translating them into Indonesian. The dataset was self-generated using the Noto Sans Batak font and character combinations, totaling 113 labels, which were processed into 64×64 grayscale images. The CNN model was designed with several convolutional and pooling layers and compiled using the Adam optimizer and categorical cross-entropy loss function. Training results achieved a validation accuracy of 98.36% and a testing accuracy of 98.12%, with respective loss values of 0.0268 and 0.0295. The system was then integrated into a web-based application built as a Progressive Web App (PWA), supporting both online transliteration and translation features. These results demonstrate that the CNN approach is highly effective in recognizing Batak Toba characters. In the future, the system can be further developed into a full sentence-level OCR, integrated into a native Android application, and expanded with datasets from real handwritten samples.

Downloads

Download data is not yet available.

References

R. Lorentius, R. Setiawan, and D. Hutabarat, “Preserving local scripts through digital recognition systems,” J. Cult. Inform., vol. 5, no. 2, pp. 45–53, 2020.

UNESCO, “Culture for sustainable development: The role of heritage in achieving SDGs,” UNESCO Rep. Cult. Dev., Paris, 2017.

J. Butar-Butar, T. Sihombing, and M. Silalahi, “Revitalization of Batak Toba script through digital media,” Indones. J. Lang. Cult., vol. 4, no. 1, pp. 22–30, 2022.

U. Kozok, Buku Tanda-Tanda: The Batak Script and Its History. Jakarta: KITLV Press, 2009.

P. Turnip, F. Sinaga, and R. Pardede, “Digital archiving of Batak script manuscripts for cultural preservation,” J. Teknol. Komput. Indones., vol. 12, no. 3, pp. 110–118, 2023.

S. Pasaribu, “Aksara Batak dan fungsinya dalam tradisi tulisan Batak Toba,” J. Kebahasaan Kesastraan Indones., vol. 7, no. 2, pp. 65–73, 2015.

R. Pratiwi, U. Hasanah, and A. Rachman, “A review of deep learning algorithms for image recognition,” Int. J. Artif. Intell. Res., vol. 5, no. 1, pp. 56–64, 2021.

M. Omori and K. Shima, “Convolutional neural network architecture for visual pattern recognition,” Procedia Comput. Sci., vol. 176, pp. 2544–2551, 2020.

R. Mulyanto, M. Fathoni, and D. Putra, “Lampung script recognition using convolutional neural networks,” J. Ilm. Teknol. Inf., vol. 19, no. 2, pp. 87–95, 2021.

R. Rikendry and M. Maharil, “Comparison of VGG-16 and ResNet-50 architectures for Lampung handwritten character recognition,” J. Comput. Vis. Image Process., vol. 8, no. 1, pp. 33–41, 2022.

D. Shelvi, R. Wijaya, and A. Rahman, “Handwritten Sundanese script recognition using convolutional neural network,” Indones. J. Artif. Intell. Data Sci., vol. 2, no. 4, pp. 121–128, 2021.

S. Handoko and D. Indahyanti, “CNN-based recognition of Bima script characters,” J. Inform. Komput. Cerdas, vol. 3, no. 1, pp. 14–22, 2024.

A. Alvin and I. Wasito, “Handwritten Javanese character recognition using deep convolutional neural network,” in Proc. 2023 Int. Conf. Inf. Technol. Innov., pp. 88–94, 2023.

A. Cahya, R. Nugraha, and D. Sari, “Digitalization of local scripts: Efforts to preserve Indonesian writing systems,” J. Kebud. Nusantara, vol. 11, no. 2, pp. 97–105, 2021.

N. Fadillah, H. Kurniawan, and R. Setiadi, “Enhancing CNN performance using elastic distortion for handwritten character recognition,” Int. J. Mach. Vis. Signal Process., vol. 9, no. 3, pp. 201–210, 2021.

F. Anhar and Y. Rahma, “Improving deep CNN models for multi-class handwritten recognition through data augmentation,” J. Comput. Sci. Appl., vol. 15, no. 2, pp. 142–151, 2023.

M. Fawwaz, S. Putri, and T. Hidayat, “Optimization of CNN with Adam and cross-entropy for character image classification,” J. Teknol. Syst. Komput., vol. 8, no. 4, pp. 155–162, 2020.