CORAA ASR: a large corpus of spontaneous and prepared speech manually validated for speech recognition in Brazilian Portuguese

dc.contributor.authorCandido Junior, Arnaldo [UNESP]
dc.contributor.authorCasanova, Edresson
dc.contributor.authorSoares, Anderson
dc.contributor.authorde Oliveira, Frederico Santos
dc.contributor.authorOliveira, Lucas
dc.contributor.authorJunior, Ricardo Corso Fernandes
dc.contributor.authorda Silva, Daniel Peixoto Pinto
dc.contributor.authorFayet, Fernando Gorgulho
dc.contributor.authorCarlotto, Bruno Baldissera
dc.contributor.authorGris, Lucas Rafael Stefanel
dc.contributor.authorAluísio, Sandra Maria
dc.contributor.institutionFederal University of Technology — Paraná (UTFPR)
dc.contributor.institutionUniversidade de São Paulo (USP)
dc.contributor.institutionFederal University of Goias
dc.contributor.institutionUniversidade Estadual Paulista (UNESP)
dc.date.accessioned2023-07-29T12:38:19Z
dc.date.available2023-07-29T12:38:19Z
dc.date.issued2022-01-01
dc.description.abstractAutomatic Speech recognition (ASR) is a complex and challenging task. In recent years, there have been significant advances in the area. In particular, for the Brazilian Portuguese (BP) language, there were around 376 h publicly available for the ASR task until the second half of 2020. With the release of new datasets in early 2021, this number increased to 574 h. The existing resources, however, are composed of audios containing only read and prepared speech. There is a lack of datasets including spontaneous speech, which are essential in several ASR applications. This paper presents CORAA (Corpus of Annotated Audios) ASR with 290 h, a publicly available dataset for ASR in BP containing validated pairs of audio-transcription. CORAA ASR also contains European Portuguese audios (4.6 h). We also present a public ASR model based on Wav2Vec 2.0 XLSR-53, fine-tuned over CORAA ASR. Our model achieved a Word Error Rate (WER) of 24.18% on CORAA ASR test set and 20.08% on Common Voice test set. When measuring the Character Error Rate (CER), we obtained 11.02% and 6.34% for CORAA ASR and Common Voice, respectively. CORAA ASR corpora were assembled to both improve ASR models in BP with phenomena from spontaneous speech and motivate young researchers to start their studies on ASR for Portuguese. All the corpora are publicly available at https://github.com/nilc-nlp/CORAA under the CC BY-NC-ND 4.0 license.en
dc.description.affiliationFederal University of Technology — Paraná (UTFPR)
dc.description.affiliationInstituto de Ciências Matemáticas e de Computação - University of São Paulo
dc.description.affiliationFederal University of Goias
dc.description.affiliationSão Paulo State University
dc.description.affiliationUnespSão Paulo State University
dc.identifierhttp://dx.doi.org/10.1007/s10579-022-09621-4
dc.identifier.citationLanguage Resources and Evaluation.
dc.identifier.doi10.1007/s10579-022-09621-4
dc.identifier.issn1572-8412
dc.identifier.issn1574-020X
dc.identifier.scopus2-s2.0-85142235932
dc.identifier.urihttp://hdl.handle.net/11449/246341
dc.language.isoeng
dc.relation.ispartofLanguage Resources and Evaluation
dc.sourceScopus
dc.subjectAutomatic speech recognition
dc.subjectBrazilian Portuguese
dc.subjectPrepared speech
dc.subjectPublic datasets
dc.subjectPublic speech corpora
dc.subjectSpontaneous speech
dc.titleCORAA ASR: a large corpus of spontaneous and prepared speech manually validated for speech recognition in Brazilian Portugueseen
dc.typeArtigo
unesp.author.orcid0000-0002-5647-0891[1]
unesp.author.orcid0000-0003-0160-7173[2]
unesp.author.orcid0000-0002-2967-6077[3]
unesp.author.orcid0000-0002-5885-6747[4]
unesp.author.orcid0000-0001-5108-2630[11]

Arquivos

Coleções