HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity

[EN] Offline Handwritten Text Recognition (HTR) systems recognize and transcribe handwritten text from scanned images into digital formats. The field has become important due to the need for document digitization and data entry automation in various industries. Accurate recognition requires large an...

Descripción completa

Detalles Bibliográficos
Autores: Neto, Arthur F. S., Bezerra, Byron L. D., Toselli, Alejandro Héctor
Tipo de recurso: artículo
Fecha de publicación:2026
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:dnet:riunet______::aa586329467f84e84ca0532747b05469
Acceso en línea:https://riunet.upv.es/handle/10251/234623
Access Level:acceso abierto
Palabra clave:Data augmentation
Data synthesis
Handwriting synthesis
Handwritten text recognition
id ES_edbb92c934867e4c7cb5d63e70965bfe
oai_identifier_str oai:dnet:riunet______::aa586329467f84e84ca0532747b05469
network_acronym_str ES
network_name_str España
repository_id_str
spelling HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data ScarcityNeto, Arthur F. S.Bezerra, Byron L. D.Toselli, Alejandro HéctorData augmentationData synthesisHandwriting synthesisHandwritten text recognition[EN] Offline Handwritten Text Recognition (HTR) systems recognize and transcribe handwritten text from scanned images into digital formats. The field has become important due to the need for document digitization and data entry automation in various industries. Accurate recognition requires large and varied datasets for training optical models, where collecting and labeling these datasets is often time-consuming and impractical. To address this challenge, data augmentation and transfer learning are commonly used. Nevertheless, these traditional methods may lead to overfitting and performance degradation when data are scarce. This work proposes integrating Conditional Generative Adversarial Networks (CGANs) for data synthesis into optical model training to improve handwriting recognition in data-scarce scenarios. To validate our proposal, we conducted a study that included: (i) an exploration to establish an optimal configuration for traditional data augmentation; and (ii) extensive experiments using seven datasets. In addition, these datasets were partitioned into training subsets to simulate diverse data-scarcity conditions. Averaged over all subsets and optical models, data synthesis achieved the highest reductions compared to the baseline trained from scratch without augmentation. It reduced the Character Error Rate (CER) by 41.1% and the Word Error Rate (WER) by 28.1%. Transfer learning achieved reductions of 34.4% in CER and 23.5% in WER. Lastly, traditional data augmentation achieved reductions of 13.8% in CER and 12.3% in WER. These findings highlight the importance of data synthesis for improving HTR systems, particularly in data-scarcity contexts.This work was supported in part by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) under Grant 001, and in part by the Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE) under Grant APQ-1216-1.03/22.Institute of Electrical and Electronics EngineersDepartamento de Sistemas Informáticos y ComputaciónEscuela Técnica Superior de Ingeniería InformáticaCentro de Investigación Pattern Recognition and Human Language TechnologyFundação de Amparo à Ciência e Tecnologia do Estado de PernambucoCoordenaçao de Aperfeiçoamento de Pessoal de Nível Superior, BrasilConselho Nacional de Desenvolvimento Científico e Tecnológico, BrasilRepositorio Institucional de la Universitat Politècnica de València Riunet20262026-01-01journal articlehttp://purl.org/coar/resource_type/c_6501VoRhttp://purl.org/coar/version/c_970fb48d4fbd8a85info:eu-repo/semantics/articleapplication/pdfhttps://riunet.upv.es/handle/10251/234623reponame:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valénciainstname:Universitat Politècnica de València (UPV)InglésengFundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco https://doi.org/10.13039/501100006162 APQ-1216-1.03%2F22open accesshttp://purl.org/coar/access_right/c_abf2Reconocimiento (by)http://creativecommons.org/licenses/by/4.0/info:eu-repo/semantics/openAccessoai:dnet:riunet______::aa586329467f84e84ca0532747b054692026-06-13T07:49:27Z
dc.title.none.fl_str_mv HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
title HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
spellingShingle HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
Neto, Arthur F. S.
Data augmentation
Data synthesis
Handwriting synthesis
Handwritten text recognition
title_short HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
title_full HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
title_fullStr HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
title_full_unstemmed HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
title_sort HTSR-Pollen: Handwritten Text Synthesis and Recognition System to Overcome Data Scarcity
dc.creator.none.fl_str_mv Neto, Arthur F. S.
Bezerra, Byron L. D.
Toselli, Alejandro Héctor
author Neto, Arthur F. S.
author_facet Neto, Arthur F. S.
Bezerra, Byron L. D.
Toselli, Alejandro Héctor
author_role author
author2 Bezerra, Byron L. D.
Toselli, Alejandro Héctor
author2_role author
author
dc.contributor.none.fl_str_mv Departamento de Sistemas Informáticos y Computación
Escuela Técnica Superior de Ingeniería Informática
Centro de Investigación Pattern Recognition and Human Language Technology
Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco
Coordenaçao de Aperfeiçoamento de Pessoal de Nível Superior, Brasil
Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brasil
Repositorio Institucional de la Universitat Politècnica de València Riunet
dc.subject.none.fl_str_mv Data augmentation
Data synthesis
Handwriting synthesis
Handwritten text recognition
topic Data augmentation
Data synthesis
Handwriting synthesis
Handwritten text recognition
description [EN] Offline Handwritten Text Recognition (HTR) systems recognize and transcribe handwritten text from scanned images into digital formats. The field has become important due to the need for document digitization and data entry automation in various industries. Accurate recognition requires large and varied datasets for training optical models, where collecting and labeling these datasets is often time-consuming and impractical. To address this challenge, data augmentation and transfer learning are commonly used. Nevertheless, these traditional methods may lead to overfitting and performance degradation when data are scarce. This work proposes integrating Conditional Generative Adversarial Networks (CGANs) for data synthesis into optical model training to improve handwriting recognition in data-scarce scenarios. To validate our proposal, we conducted a study that included: (i) an exploration to establish an optimal configuration for traditional data augmentation; and (ii) extensive experiments using seven datasets. In addition, these datasets were partitioned into training subsets to simulate diverse data-scarcity conditions. Averaged over all subsets and optical models, data synthesis achieved the highest reductions compared to the baseline trained from scratch without augmentation. It reduced the Character Error Rate (CER) by 41.1% and the Word Error Rate (WER) by 28.1%. Transfer learning achieved reductions of 34.4% in CER and 23.5% in WER. Lastly, traditional data augmentation achieved reductions of 13.8% in CER and 12.3% in WER. These findings highlight the importance of data synthesis for improving HTR systems, particularly in data-scarcity contexts.
publishDate 2026
dc.date.none.fl_str_mv 2026
2026-01-01
dc.type.none.fl_str_mv journal article
http://purl.org/coar/resource_type/c_6501
VoR
http://purl.org/coar/version/c_970fb48d4fbd8a85
dc.type.openaire.fl_str_mv info:eu-repo/semantics/article
format article
dc.identifier.none.fl_str_mv https://riunet.upv.es/handle/10251/234623
url https://riunet.upv.es/handle/10251/234623
dc.language.none.fl_str_mv Inglés
eng
language_invalid_str_mv Inglés
language eng
dc.relation.none.fl_str_mv Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco https://doi.org/10.13039/501100006162 APQ-1216-1.03%2F22
dc.rights.none.fl_str_mv open access
http://purl.org/coar/access_right/c_abf2
Reconocimiento (by)
http://creativecommons.org/licenses/by/4.0/
dc.rights.openaire.fl_str_mv info:eu-repo/semantics/openAccess
rights_invalid_str_mv open access
http://purl.org/coar/access_right/c_abf2
Reconocimiento (by)
http://creativecommons.org/licenses/by/4.0/
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.publisher.none.fl_str_mv Institute of Electrical and Electronics Engineers
publisher.none.fl_str_mv Institute of Electrical and Electronics Engineers
dc.source.none.fl_str_mv reponame:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
instname:Universitat Politècnica de València (UPV)
instname_str Universitat Politècnica de València (UPV)
reponame_str RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
collection RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
repository.name.fl_str_mv
repository.mail.fl_str_mv
_version_ 1869423530269999104
score 15.81155

Similares en LA Referencia