Pengenalan Sekuens Kata Aksara Jawa Tanpa Segmentasi Menggunakan Arsitektur CRNN dan CTC
DOI:
https://doi.org/10.29407/trcqtt87Abstract
Terdapat penurunan kemampuan masyarakat dalam membaca Aksara Jawa, yang sebagian besar disebabkan oleh kompleksitas aturan penulisan Sandhangan dan Pasangan yang posisinya dapat bertumpuk secara vertikal. Metode Optical Character Recognition (OCR) konvensional yang berbasis segmentasi karakter sering mengalami kegagalan karena pemotongan karakter justru merusak informasi visual dari aksara yang bertumpuk. Penelitian ini mengusulkan pendekatan segmentation-free menggunakan arsitektur Convolutional Recurrent Neural Network (CRNN) yang dipadukan dengan Connectionist Temporal Classification (CTC) Loss. Model dilatih menggunakan lebih dari 15.000 citra dataset sintetis kata Aksara Jawa yang telah melalui proses augmentasi. Hasil pengujian dievaluasi menggunakan metrik Character Error Rate (CER) dan Word Error Rate (WER) berdasarkan jarak Levenshtein. Penelitian ini membuktikan bahwa pendekatan pengenalan sekuens kata secara utuh mampu menerjemahkan Aksara Jawa dengan tingkat akurasi karakter mencapai 99,71%, sehingga dapat menjadi solusi digitalisasi aksara tradisional yang efektif.
Keywords:
Aksara Jawa, CRNN, CTC, Deep Learning, Transliterasi##plugins.themes.default.displayStats.downloads##
References
[1] E. K. Aribowo, “DIGITALISASI AKSARA JAWA DAN
PEMANFAATANNYA SEBAGAI MEDIA PEMBELAJARAN BAGI
MUSYAWARAH GURU MATA PELAJARAN BAHASA JAWA SMP
KABUPATEN KLATEN,” Warta LPM, vol. 21, no. 2, pp. 59–70, Jun. 2018, doi:
10.23917/warta.v21i2.5620.
[2] Erviana, “Pemanfaatan Teknologi Optical Character Recognition (Untuk
Mengenali Alfabet Aksara Jawa Berbasis Android Dengan Algoritma
Convulutional Neural Network),” UNIVERSITAS BUDDHI DHARMA
TANGERANG, 2023. [Online]. Available:
https://repositori.buddhidharma.ac.id/1988/2/COVER - BAB 3.pdf
[3] Y. Sugianela and N. Suciati, “CHARACTER IMAGE SEGMENTATION OF
JAVANESE SCRIPT USING CONNECTED COMPONENT METHOD,”
Jurnal Ilmu Komputer dan Informasi, vol. 12, no. 2, pp. 67–74, Jul. 2019, doi:
10.21609/jiki.v12i2.677.
[4] A. Akoushideh, A. Ranjkesh Rashtehroudi, and A. Shahbahrami, “Persian/Arabic
Scene Text Recognition With Convolutional Recurrent Neural Network,” IET
Smart Cities, vol. 7, no. 1, Jan. 2025, doi: 10.1049/smc2.70001.
[5] M. Nagaraju, P. Chawla, and N. Kumar, “Performance improvement of Deep
Learning Models using image augmentation techniques,” Multimed. Tools Appl.,
vol. 81, no. 7, pp. 9177–9200, Mar. 2022, doi: 10.1007/s11042-021-11869-x.
[6] R. Siringoringo, J. Jamaluddin, R. Perangin-angin, E. J. G. Harianja, G.
Lumbantoruan, and E. N. Purba, “MODEL BIDIRECTIONAL LSTM UNTUK
PEMROSESAN SEKUENSIAL DATA TEKS SPAM,” METHOMIKA Jurnal
Manajemen Informatika dan Komputerisasi Akuntansi, vol. 7, no. 2, pp. 265–271,
Oct. 2023, doi: 10.46880/jmika.Vol7No2.pp265-271.
[7] Z. Zhang et al., “Self-Distillation Regularized Connectionist Temporal
Classification Loss for Text Recognition: A Simple Yet Effective Approach,”
Proceedings of the AAAI Conference on Artificial Intelligence, vol. 38, no. 7, pp.
7441–7449, Mar. 2024, doi: 10.1609/aaai.v38i7.28575.
[8] R. P. Hati, “EVALUASI KINERJA SPEECH-TO-TEXT DENGAN REDUKSI
KEBISINGAN SPECTRAL GATING DAN WIENER FILTERING PADA
AUDIO BAHASA LAMPUNG,” UNIVERSITAS LAMPUNG, 2025. Accessed:
Jun. 04, 2026. [Online]. Available: http://digilib.unila.ac.id/id/eprint/88775
[9] Muhammad Nashiruddin, F. D. Noor Praditya, Agiel Faiz Mufazzal, and A.
Ardiansyah, “Implementation of Tesseract-based OCR for UMKLA student card
data extraction,” JKTI Jurnal Keilmuan Teknologi Informasi, vol. 1, no. 1, pp. 11–
14, Jun. 2025, doi: 10.61902/jkti.v1i1.1688.
[10] L. Abdiansah, S. Sumarno, A. Eviyanti, and N. L. Azizah, “Penerapan Algoritma
Convolutional Neural Networks untuk Pengenalan Tulisan Tangan Aksara Jawa,”
MALCOM: Indonesian Journal of Machine Learning and Computer Science, vol.
5, no. 2, pp. 496–504, Mar. 2025, doi: 10.57152/malcom.v5i2.1814.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Copyright on any article is retained by the author(s).
- The author grants the journal, right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work’s authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal’s published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
- The article and any associated published material is distributed under the Creative Commons Attribution-ShareAlike 4.0 International License
