Abstract
This paper proposes a method for correcting character-recognition errors in Japanese handwritten answers to writing-type questions from exercise books. We created a model to correct character-recognition errors by fine-tuning the text-to-text-transfer-transformer (T5) using pairs of automatically recognized data from handwritten answers and their manual corrections. The data comprised handwritten Japanese answers from 185 junior high school students to writing-type questions in a Japanese language task. In addition, we augmented the training data using the five best results of the character-recognition model with confidence scores to learn additional patterns of recognition errors. The experimental results revealed that the answers corrected by the proposed method were closer to the actual answers than those before the correction and data augmentation was effective for the correction model.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
- 2.
- 3.
The vertical-writing recognition model did not obtain the best five results for all samples. It output less than five results for some of them because the candidates with confidence scores below a specific threshold were omitted from the system output.
- 4.
The single-character-recognition model did not output the best five results for all samples. Some of them had less than five results because the candidates with confidence scores below the specified threshold were omitted from the system outputs.
- 5.
We tried multiplications with 1/100, 1/10, and 1.
- 6.
- 7.
- 8.
There are no words like “目由” and “買えふ” in Japanese.
- 9.
The only comprehensible parts of this sentence are “to buy books” and “I should read.”.
- 10.
The system output means that the same animals and plants are found on both sides of Africa and South America and the reference gold data states that, in the East, where the idea is to link body and spirit, it is very important to maintain a good posture.
References
Atienza, R.: Data augmentation for scene text recognition. In: 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), pp. 1561–1570 (2021). https://doi.org/10.1109/ICCVW54120.2021.00181
Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: NAACL-HLT2019, pp. 4171–4186 (2019)
Hayashi, T., Gyohten, K., Ohki, H., Takami, T.: A study of data augmentation for handwritten character recognition using deep learning. In: 2018 16th International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 552–557 (2018). https://doi.org/10.1109/ICFHR-2018.2018.00102
Katinskaia, A., Yangarber, R.: Grammatical error correction for sentence-level assessment in language learning. In: Proceedings of the 18th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2023). Association for Computational Linguistics (2023)
Nagata, M.: Japanese OCR error correction using character shape similarity and statistical language model. In: COLING 1998 Volume 2: The 17th International Conference on Computational Linguistics (1998). https://aclanthology.org/C98-2147
Nakamura, A., Li, S., Tamura, K., Yoshinaga, N.: Error correction of neural speech recognition considering pre- and post-speech as context (Zengo no hatsuwa wo bunmyaku tosite kouryosuru neural onsei ninshiki ayamariteisei). J. Inf. Process. Soc. Jpn. 1–7 (2022)
Nguyen, Q.D., Phan, N.M., Krömer, P., Le, D.A.: An efficient unsupervised approach for OCR error correction of Vietnamese OCR text. IEEE Access 11, 58406–58421 (2023). https://doi.org/10.1109/ACCESS.2023.3283340
Papineni, K., Roukos, S., Ward, T., Zhu, W.J.: BLEU: a method for automatic evaluation of machine translation. In: ACL, pp. 311–318 (2002)
Qorib, M.R., Ng, H.T.: Grammatical error correction: are we there yet? In: Proceedings of the 29th International Conference on Computational Linguistics, Gyeongju, Republic of Korea. International Committee on Computational Linguistics (2022)
Raffel, C., et al.: Exploring the limits of transfer learning with a unified text-to-text transformer. J. Mach. Learn. Res. 21(140), 1–67 (2020)
Sakamoto, K., et al.: Correction of word errors in OCR of contract documents using tree edit distance for the considering the radicals of a kanji character (Keiyakusyo OCR no tango ayamariteisei ni okeru kanji no henboukankyaku wo kouryosita kihensyuukyori no kentou). The Association for Natural Language Processing, pp. 137–140 (2020)
Shahgir, H.S., Sayeed, K.S.: Bangla grammatical error detection using T5 transformer model. arXiv:2303.10612 [cs.CL] (2023)
Soma, N., Kajiura, T., Takahashi, M., Kuramitsu, K.: Applying error correction models in code with additional pre-training to large language model (Daikibo gengo model heno tsuika jizengakusyu niyoru ayamariteisei model no code heno tekiyou). DEIM Forum 2023, pp. 1b–5–4 (2023)
Spruck, A., Hawesch, M., Maier, A., Riess, C., Seiler, J., Kaup, A.: 3D rendering framework for data augmentation in optical character recognition. arXiv:2209.14970 (2022)
Takeuchi, K., Matsumoto, Y.: OCR error correction using stochastic language models (Tokeiteki gengo model wo mochiita OCR ayamari system no kouchiku). J. Inf. Process. Soc. Jpn. 40(6), 2679–2689 (1999)
Xie, S., Matsumoto, A.: Accuracy improving on digitization of modern Japanese documents with pre-trained BERT model (Nihongo BERT model ni yoru kindaibun no ayamariteisei). The Association for Natural Language Processing, pp. 1616–1620 (2023)
Zhang, Y., Kamigaito, H., Okumura, M.: Bidirectional transformer reranker for grammatical error correction. In: Findings of the Association for Computational Linguistics: ACL 2023. Association for Computational Linguistics, Toronto, Canada (2023)
Acknowledgements
This work was supported by JSPS KAKENHI Grant Numbers 23H03511 and 24H00738. The answers were collected with the approval of the University’s Ethics Review Committee for Research Involving Human Subjects (No. 220707-04111). We would like to thank Toshihiko Horie from Wacom Co., Ltd., who gave us the text data set of questions.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Suzuki, R. et al. (2024). Error Correction of Japanese Character-Recognition in Answers to Writing-Type Questions Using T5. In: Sfikas, G., Retsinas, G. (eds) Document Analysis Systems. DAS 2024. Lecture Notes in Computer Science, vol 14994. Springer, Cham. https://doi.org/10.1007/978-3-031-70442-0_14
Download citation
DOI: https://doi.org/10.1007/978-3-031-70442-0_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-70441-3
Online ISBN: 978-3-031-70442-0
eBook Packages: Computer ScienceComputer Science (R0)
