CORAA ASR: a large corpus of spontaneous and prepared speech manually validated for speech recognition in Brazilian Portuguese
dc.contributor.author | Candido Junior, Arnaldo [UNESP] | |
dc.contributor.author | Casanova, Edresson | |
dc.contributor.author | Soares, Anderson | |
dc.contributor.author | de Oliveira, Frederico Santos | |
dc.contributor.author | Oliveira, Lucas | |
dc.contributor.author | Junior, Ricardo Corso Fernandes | |
dc.contributor.author | da Silva, Daniel Peixoto Pinto | |
dc.contributor.author | Fayet, Fernando Gorgulho | |
dc.contributor.author | Carlotto, Bruno Baldissera | |
dc.contributor.author | Gris, Lucas Rafael Stefanel | |
dc.contributor.author | Aluísio, Sandra Maria | |
dc.contributor.institution | Federal University of Technology — Paraná (UTFPR) | |
dc.contributor.institution | Universidade de São Paulo (USP) | |
dc.contributor.institution | Federal University of Goias | |
dc.contributor.institution | Universidade Estadual Paulista (UNESP) | |
dc.date.accessioned | 2023-07-29T12:38:19Z | |
dc.date.available | 2023-07-29T12:38:19Z | |
dc.date.issued | 2022-01-01 | |
dc.description.abstract | Automatic 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.affiliation | Federal University of Technology — Paraná (UTFPR) | |
dc.description.affiliation | Instituto de Ciências Matemáticas e de Computação - University of São Paulo | |
dc.description.affiliation | Federal University of Goias | |
dc.description.affiliation | São Paulo State University | |
dc.description.affiliationUnesp | São Paulo State University | |
dc.identifier | http://dx.doi.org/10.1007/s10579-022-09621-4 | |
dc.identifier.citation | Language Resources and Evaluation. | |
dc.identifier.doi | 10.1007/s10579-022-09621-4 | |
dc.identifier.issn | 1572-8412 | |
dc.identifier.issn | 1574-020X | |
dc.identifier.scopus | 2-s2.0-85142235932 | |
dc.identifier.uri | http://hdl.handle.net/11449/246341 | |
dc.language.iso | eng | |
dc.relation.ispartof | Language Resources and Evaluation | |
dc.source | Scopus | |
dc.subject | Automatic speech recognition | |
dc.subject | Brazilian Portuguese | |
dc.subject | Prepared speech | |
dc.subject | Public datasets | |
dc.subject | Public speech corpora | |
dc.subject | Spontaneous speech | |
dc.title | CORAA ASR: a large corpus of spontaneous and prepared speech manually validated for speech recognition in Brazilian Portuguese | en |
dc.type | Artigo | |
unesp.author.orcid | 0000-0002-5647-0891[1] | |
unesp.author.orcid | 0000-0003-0160-7173[2] | |
unesp.author.orcid | 0000-0002-2967-6077[3] | |
unesp.author.orcid | 0000-0002-5885-6747[4] | |
unesp.author.orcid | 0000-0001-5108-2630[11] |