Enhancing Lontara Script Handwritten Recognition with Zoning and Convolutional Neural Networks

Sri Wulan Dari, Zahir Zainuddin, Mukarramah Yusuf, Herviana

2025

Abstract

This paper addresses the challenge of improving handwritten recognition of Lontara script, particularly for 23 characters and an additional 5 characters with the O diacritic. Recognition errors often occur due to the high visual similarity between characters and the diversity in handwriting styles, which remain significant barriers in the existing literature. Despite advances with contour-based features and sliding window methods, confusion between visually similar characters such as 'Ta' and the diacritic 'O' remains unresolved. To fill this gap, this study introduces an integrated approach that combines Zoning for enhanced feature extraction with Convolutional Neural Networks (CNN) for classification. The proposed method overcomes these challenges by capturing distinct localized features, which are crucial for accurate recognition, improving the classification accuracy by 21%. This improvement significantly enhances the model's ability to differentiate similar characters, thus contributing to more reliable handwritten Lontara character recognition.

Paper Citation

in Harvard Style

Dari S., Zainuddin Z., Yusuf M. and Herviana. (2025). Enhancing Lontara Script Handwritten Recognition with Zoning and Convolutional Neural Networks. In Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH; ISBN 978-989-758-784-9, SciTePress, pages 84-89. DOI: 10.5220/0014276400004928

in Bibtex Style

@conference{ritech25,
author={Sri Wulan Dari and Zahir Zainuddin and Mukarramah Yusuf and Herviana},
title={Enhancing Lontara Script Handwritten Recognition with Zoning and Convolutional Neural Networks},
booktitle={Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH},
year={2025},
pages={84-89},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0014276400004928},
isbn={978-989-758-784-9},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Innovations in Information and Engineering Technology - Volume 1: RITECH
TI - Enhancing Lontara Script Handwritten Recognition with Zoning and Convolutional Neural Networks
SN - 978-989-758-784-9
AU - Dari S.
AU - Zainuddin Z.
AU - Yusuf M.
AU - Herviana.
PY - 2025
SP - 84
EP - 89
DO - 10.5220/0014276400004928
PB - SciTePress