HTR and assessment of a handwritten Basque text corpus from primary education
Article Sidebar
Main Article Content
DOI:
https://doi.org/10.17979/digilec.2026.13.13509Abstract
The Primary Education (EP) test for the Diagnostic Evaluation (ED) of the Basque Educational System within the specific domain and field of language are collected as handwritten textual productions. If these texts were digitized, they could be analyzed both manually by experts and automatically with Natural Language Processing (NLP) techniques. The principal objective of this study is to develop a Handwritten Text Recognition (HTR) model for the digitization of handwritten texts in Basque produced by 4th grade Primary Education students, as well as to create a digitized corpus of these texts. Although Handwritten Text Recognition (HTR) models in Basque exist for older students, a model was developed specifically adapted to the young students’ writing characteristics. A total of 870 handwritten texts provided by the Basque Institute of Educational Evaluation and Research (ISEI-IVEI) were collected, and the Transkribus platform was used, applying advanced Artificial Intelligence (AI) techniques to train the models. The evaluation process was carried out manually, and a qualitative analysis of the model’s errors was conducted using a coding system. The results show an error rate between 5 % and 7 %, representing an advance in the digitization of children’s handwritten texts. Finally, both the Handwritten Text Recognition (HTR) model and the corpus are expected to be made available to the scientific and educational community, facilitating future research in this area.
Keywords:
Article Details
References
Adak, C., Chaudhuri, B. B. y Blumenstein, M. (2017). Legibility and aesthetic analysis of handwriting. In 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR) (Vol. 1, pp. 175-182). IEEE. doi: 10.1109/ICDAR.2017.37.
Berninger, V. W. y Richards, T. L. (2010). Inter-relationships among behavioral markers, genes, brain, and treatment in dyslexia and dysgraphia. Future Neurology, 5(4), 597–617. https://doi.org/10.2217/fnl.10.22
Butler, T. S. (2003). Human interaction with digital ink: legibility measurement and structural analysis [Tesis de Doctorado]. University of Hertfordshire. http://hdl.handle.net/2299/14158
Dinehart, L. H. (2015). Handwriting in early childhood education: Current research and future implications. Journal of Early Childhood Literacy, 15, 97–118. DOI: https://doi.org/10.1177/1468798414522825
Durrant, P. (2022). Studying children's writing development with a corpus. Applied Corpus Linguistics, 2(3), 100026. https://doi.org/10.1016/j.acorp.2022.100026
Fischer, A., Frinken, V., Bunke, H. y Fornés, A. (2012). Lexicon-free handwritten word spotting using character HMMs. Pattern Recognition Letters, 33(7), 934–942. https://doi.org/10.1016/j.patrec.2011.09.009
Fogel, Y., Rosenblum, S. y Barnett, A. L. (2022). Handwriting legibility across different writing tasks in school-aged children. Hong Kong journal of occupational therapy: HKJOT, 35(1), 44–51. https://doi.org/10.1177/15691861221075709
González, R. M. (2003). Propuesta de intervención en los procesos cognitivos y estructuras textuales en niños con DAE. Psicothema, 15(3), 458-463.
Government of South Australia, Department of Education and Children’s Services (2007). Handwriting in the South Australian Curriculum (2nd ed.; G. Groves, Ed.). Hyde Park Press. https://www.australianschoolfonts.com.au/assets/resources/handwriting-2nd-ed-2007-sa.pdf
Graves, A., Liwicki, M., Fernández, S., Bertolami, R., Bunke, H. y Schmidhuber, J. (2009). A novel connectionist system for unconstrained handwriting recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 31(5), 855–868. https://doi.org/10.1109/TPAMI.2008.137
Ibarra, I. (2016). Esku-idazketa eta testu eleanitzen arteko loturak. [Tesis de Doctorado]. Universidad del País Vasco/Euskal Herriko Unibertsitatea (UPV/EHU). https://addi.ehu.eus/bitstream/handle/10810/26312/TESIS_IBARRA_LIZUNDIA_IRUNE.pdf?sequence=1&isAllowed=y
Ibarra, I., Ortube, M., eta Muxika, I. (2020). Letra xeheak, bai! Booktegi. Recuperado de https://www.booktegi.eus/liburuak/letra-xeheakbai/
Ingle, R., Fujii, Y., Deselaers, T., Baccash, J. y Popat, A. (2019). A scalable handwritten text recognition system. Proceedings of ICDAR 2019, 17–24. https://doi.org/10.48550/arXiv.1904.09150
Instituto Vasco de Evaluación e Investigación Educativa (ISEI-IVEI). (2025). Evaluación diagnóstica de mitad de etapa. Competencia en comunicación lingüística. Gobierno Vasco. https://isei-ivei.euskadi.eus/es/
Jaipukar, P.B., Pal V.S., Varade, S.M. y Mendhule, S.P. (2025). Interpreting Handwriting Text Recognition. International Journal of Progressive Research in Engineering Management and Science (IJPREMS) (Int Peer Reviewed Journal), 5(8), 1745-1748. https://www.ijprems.com/uploadedfiles/paper//issue_8_august_2025/43592/final/fin_ijprems1756717555.pdf
Kahle, P., Colutto, S., Hackl, G. y Mühlberger, G. (2017). Transkribus—A service platform for transcription, recognition and retrieval of historical documents. Proceedings of ICDAR 2017, 19–24. https://doi.org/10.1109/ICDAR.2017.307
Ley Orgánica 2/2006, de 3 de mayo, de Educación. (BOE núm. 106, de 4 de mayo de 2006, pp. 1-112).
Mahadevkar, S., Patil, S. y Kotecha, K. (2024). Enhancement of handwritten text recognition using AI-based hybrid approach. MethodsX, 12, 1-11. DOI: https://doi.org/10.1016/j.mex.2024.102654
Muehlberger, G., Seaward, L., Terras, M., Ares Oliveira, S., Bosch, V., Bryan, M., Colutto, S., Déjean, H., Diem, M., Fiel, S., Gatos, B., Greinoecker, A., Grüning, T., Hackl, G., Haukkovaara, V., Heyer, G., Hirvonen, L., Hodel, T., Jokinen, M., ... Zagoris, K. (2019). Transforming scholarship in the archives through handwritten text recognition: Transkribus as a case study. Journal of Documentation, 75(5), 954-976. https://www.emerald.com/insight/content/doi/10.1108/jd-07-2018-0114/full/html
Nisa, H., Thom, J., Ciesielski, V. y Tennakoon, R. (2023). Student Messy Handwritten Dataset (SMHD). RMIT University. Dataset. https://doi.org/10.25439/rmt.24312715.v1
Nockels, J., Gooding, P., Ames, S., & Terras, M. (2022). Understanding the application of handwritten text recognition technology in heritage contexts: a systematic review of Transkribus in published research. Archival science, 22(3), 367–392. https://doi.org/10.1007/s10502-022-09397-0
Rakesh, S., Kushal Reddy, P., V. Prashanth, V. y Srinath Reddy, K. (2024). Handwritten text recognition using deep learning techniques: A survey. MATEC Web of Conferences, 392(01126), 1-8. DOI: https://doi.org/10.1051/matecconf/202439201126
Stokes, P., y Kiessling, B. (2024). Sharing Data for Handwritten Text Recognition (HTR). Digital Humanities in Practice. ffhal-04444641ff
Transkribus. (2025). Character Error Rate and Learning Curve. Transkribus Help Center. https://help.transkribus.org/character-error-rate-and-learning-curve,