Robust Balinese Script Detection for Transliteration Using YOLOv8

Hans Juwiantho; Jonathan Eka Saputra

doi:10.34148/teknika.v15i1.1466

Authors

Hans Juwiantho Informatics Department, Petra Christian University, Surabaya, East Java, Indonesia
Jonathan Eka Saputra Informatics Department, Petra Christian University, Surabaya, East Java, Indonesia

DOI:

https://doi.org/10.34148/teknika.v15i1.1466

Keywords:

Balinese Script Detection, YOLOv8, Hyperparameter Tuning, Robustness Evaluation

Abstract

YOLOv8-based detection system was developed to support automatic transliteration of Balinese script by addressing challenges related to small character size, visual similarity, and varying image quality. The study evaluates detection performance using a Balinese script dataset consisting of four dataset variations, Original, Grayscale, Augmented, and book-excluded dataset, designed to analyze the impact of preprocessing and data variation on model accuracy. The Original and Grayscale datasets demonstrating superior performance by achieving mAP@50 values above 0.93. Hyperparameter tuning experiments involving optimizer selection, epoch variation, and model scale indicated that the YOLOv8x variant with the auto optimizer and a fine-tuning strategy delivered the highest performance, reaching an mAP@50 of 0.9405 while maintaining balanced precision and recall. Confusion matrix analysis showed that most detection errors occurred between visually similar characters. Robustness evaluation further revealed that detection performance remains reliable when image clarity moderate, rotation does not exceed ±3°, shear remains within ±2°, and character bounding boxes occupy approximately at least 1% of the total image dimension. These findings contribute the importance of dataset preparation, model configuration, and robustness constraints to achieve accurate Balinese script detection for real-world transliteration applications.

Downloads

Download data is not yet available.

References

[1] K. R. Hardiyanti, I. P. Purnami, and I. B. P. M. Aryana, “Meningkatkan Kemampuan Menulis Aksara Bali Menggunakan Aplikasi Transliterasi Aksara Bali Kelas XI MIA 1,” Jurnal Pendidikan Bahasa Bali Undiksha, vol. 9, no. 1, pp. 1–10, Mar. 2022, doi: 10.23887/jpbb.v9i1.35934.

[2] Governor, “ Peraturan Daerah (Perda) Provinsi Bali Nomor 1 Tahun 2018 tentang Bahasa, Aksara, dan Sastra Bali.” Accessed: Feb. 10, 2026. [Online]. Available: https://peraturan.bpk.go.id/Details/85710/perda-prov-bali-no-1-tahun-2018

[3] Governor, “Bali Governor Regulation number 80 of 2018 About Protection and Use Of Bali, Aksara, And Literature as well as the Implementation of the Bali language,” https://jdih.baliprov.go.id/dokumen-hukum/b89a7884-30d0-4826-900b-9b53ecb112c3. Accessed: Feb. 10, 2026. [Online]. Available: https://jdih.baliprov.go.id/dokumen-hukum/b89a7884-30d0-4826-900b-9b53ecb112c3

[4] D. A. Putu et al., “Pengembangan Metode Pemilahan Suku Kata Pada Transliterasi Teks Latin Ke Aksara Bali Berbasis Finite State Machine Dengan Font Noto Serif Balinese,” Jurnal Ilmu Komputer Indonesia (JIK), vol. 6, no. 2, 2021, doi: https://doi.org/10.23887/jik.v6i2.3659.

[5] C. Oka Birawidya, G. Indrawan, and I. A. Gunadi, “Pengembangan Transliterasi Teks Aksara Bali Ke Latin Menggunakan Finite State Machine,” Jurnal Ilmu Komputer Indonesia (JIK), vol. 7, no. 1, 2022, doi: https://doi.org/10.23887/jik.v7i1.3851.

[6] I. G. N. A. C. P. I Gusti Ngurah, I. K. A. A. Putra, I. B. G. Dwidasmara, M. Widiartha, N. A. S. ER, and I. P. G. H. Suputra, “Implementasi Metode Convolutional Neural Network Pada Pengenalan Aksara Bali Berbasis Game Edukasi,” SINTECH (Science and Information Technology) Journal, vol. 6, no. 1, pp. 1–15, Apr. 2023, doi: 10.31598/sintechjournal.v6i1.1298.

[7] K. A. S. Wijaya and W. Widyastuti, “Recognition of Balinese letters with convolutional neural network,” 2022, p. 030009. doi: 10.1063/5.0103942.

[8] D. A. S. Dewi, D. M. S. Arsa, G. A. A. Putri, and N. L. P. L. S. Setiawati, “Ensembling Deep Convolutional Neural Neworks For Balinese Handwritten Character Recognition,” ASEAN Engineering Journal, vol. 13, no. 3, pp. 133–139, Aug. 2023, doi: 10.11113/aej.v13.19582.

[9] A. A. Pratama, M. D. Sulistiyo, and A. F. Ihsan, “Balinese Script Handwriting Recognition Using Faster R-CNN,” Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi), vol. 7, no. 6, pp. 1268–1275, Nov. 2023, doi: 10.29207/resti.v7i6.5176.

[10] M. W. A. Kesiman and K. T. Dermawan, “AKSALont: Automatic transliteration application for Balinese palm leaf manuscripts with LSTM Model,” Jurnal Teknologi dan Sistem Komputer, vol. 9, no. 3, pp. 142–149, Jul. 2021, doi: 10.14710/jtsiskom.2021.13969.

[11] Muhammad Husni Wahid and Erik Iman Heri Ujianto, “Efficient Pattern Recognition of Sundanese Script Variants Using CNN,” Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi), vol. 8, no. 6, pp. 808–818, Dec. 2024, doi: 10.29207/resti.v8i6.6122.

[12] A. A. Handoko, M. A. Rosid, and U. Indahyanti, “Implementasi Convolutional Neural Network (CNN) Untuk Pengenalan Tulisan Tangan Aksara Bima,” SMATIKA JURNAL, vol. 14, no. 01, pp. 96–110, Jul. 2024, doi: 10.32664/smatika.v14i01.1196.

[13] I. M. D. R. Mudiarta et al., “Balinese character recognition on mobile application based on tesseract open source OCR engine,” J. Phys. Conf. Ser., vol. 1516, no. 1, p. 012017, Apr. 2020, doi: 10.1088/1742-6596/1516/1/012017.

[14] Ilham Fikriansyah, “Aksara Bali Lengkap Beserta Arti dan Aturan Menulisnya.” Accessed: Feb. 10, 2026. [Online]. Available: https://www.detik.com/bali/budaya/d-6366936/aksara-bali-lengkap-beserta-arti-dan-aturan-menulisnya

[15] S. Hudaa, “Transliterasi, Serapan, dan Padanan Kata: Upaya Pemutakhiran Istilah dalam Bahasa Indonesia,” SeBaSa, vol. 2, no. 1, pp. 1–6, May 2019, doi: 10.29408/sbs.v2i1.1346.

[16] I. K. Paramarta, A. M. Mbete, I. B. P. Yadnya, and A. A. P. Putra, “Transliteration From Latin Into Balinese Script (Aksarabali) Using Computerized Program Of Bali Simbar,” e-Journal of Linguistics, 2015, [Online]. Available: https://ojs.unud.ac.id/index.php/eol/article/view/11972

[17] R. Sapkota, D. Ahmed, and M. Karkee, “Comparing YOLOv8 and Mask R-CNN for instance segmentation in complex orchard environments,” Oct. 2025, doi: 10.1016/j.aiia.2024.07.001.

[18] J. Terven, D.-M. Córdova-Esparza, and J.-A. Romero-González, “A Comprehensive Review of YOLO Architectures in Computer Vision: From YOLOv1 to YOLOv8 and YOLO-NAS,” Mach. Learn. Knowl. Extr., vol. 5, no. 4, pp. 1680–1716, Nov. 2023, doi: 10.3390/make5040083.

[19] T. M. Al-Hasan, V. Bonnefille, and F. Bensaali, “Enhanced YOLOv8-Based System for Automatic Number Plate Recognition,” Technologies (Basel)., vol. 12, no. 9, p. 164, Sep. 2024, doi: 10.3390/technologies12090164.

[20] L. Gumirang, A. Pramono, and G. Triyono, “Integration of Yolov8 and OCR As E-KTP Data Extraction and Validation Solution for Digital Administration Automation,” Eduvest - Journal of Universal Studies, vol. 5, pp. 13954–13965, Feb. 2025, doi: 10.59188/eduvest.v5i11.52365.

[21] R. Laroca, L. A. Zanlorensi, G. R. Gonçalves, E. Todt, W. R. Schwartz, and D. Menotti, “An efficient and layout‐independent automatic license plate recognition system based on the YOLO detector,” IET Intelligent Transport Systems, vol. 15, no. 4, pp. 483–503, Apr. 2021, doi: 10.1049/itr2.12030.

[22] N. P. Sutramiani, N. Suciati, and D. Siahaan, “MAT-AGCA: Multi Augmentation Technique on small dataset for Balinese character recognition using Convolutional Neural Network,” ICT Express, vol. 7, no. 4, pp. 521–529, Dec. 2021, doi: 10.1016/j.icte.2021.04.005.

[23] Z. Liu, R. Song, K. Li, and Y. Li, “From Detection to Understanding: A Systematic Survey of Deep Learning for Scene Text Processing,” Applied Sciences, vol. 15, no. 17, p. 9247, Aug. 2025, doi: 10.3390/app15179247.

[24] R. Harizi, R. Walha, and F. Drira, “Towards complex scene text reading with selective region proposal and two-stage deep reinforcement learning,” Appl. Soft Comput., vol. 170, p. 112701, Feb. 2025, doi: 10.1016/j.asoc.2025.112701.