Logo do repositório
 
Publicação

Interpretability analysis of deep models for COVID-19 detection

2024Artigo científico Acesso aberto
Ver registo simples
dc.contributor.authorSilva, Daniel
dc.contributor.authorCasanova, Edresson
dc.contributor.authorGris, Lucas
dc.contributor.authorGauy, Gustavo
dc.contributor.authorJunior, Arnaldo
dc.contributor.authorFinger, Marcelo
dc.contributor.authorSvartman, Flaviane
dc.contributor.authorMedeiros, Beatriz
dc.contributor.authorMartins, Marcus
dc.contributor.authorAluísio, Sandra
dc.contributor.authorBerti, Larissa
dc.contributor.authorTeixeira, João Paulo
dc.date.accessioned2026-05-18T16:25:22Z
dc.date.available2026-05-18T16:25:22Z
dc.date.issued2024
dc.description.abstract<jats:p>During the coronavirus disease 2019 (COVID-19) pandemic, various research disciplines collaborated to address the impacts of severe acute respiratory syndrome coronavirus-2 infections. This paper presents an interpretability analysis of a convolutional neural network-based model designed for COVID-19 detection using audio data. We explore the input features that play a crucial role in the model&amp;rsquo;s decision-making process, including spectrograms, fundamental frequency (F0), F0 standard deviation, sex, and age. Subsequently, we examine the model&amp;rsquo;s decision patterns by generating heat maps to visualize its focus during the decision-making process. Emphasizing an explainable artificial intelligence approach, our findings demonstrate that the examined models can make unbiased decisions even in the presence of noise in training set audios, provided appropriate preprocessing steps are undertaken. Our top-performing model achieves a detection accuracy of 94.44%. Our analysis indicates that the analyzed models prioritize high-energy areas in spectrograms during the decision process, particularly focusing on high-energy regions associated with prosodic domains, while also effectively utilizing F0 for COVID-19 detection.</jats:p>
dc.description.sponsorshipThis work was supported by FAPESP grants 2022/16374-6 (MMG), 2020/06443-5 (SPIRA), and 2023/00488-5 Artificial Intelligence in Health Interpretability of deep models for COVID-19 Volume X Issue X (2024) 11 doi: 10.36922/aih.2992 (SPIRA-BM) and by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001.
dc.identifier.citationSilva; Daniel; Casanova, Edresson; Gris, Lucas; Gauy, Gustavo; Junior, Arnaldo; Finger, Marcelo; Svartman, Flaviane; Medeiros, Beatriz; Martins, Marcus; Aluísio, Sandra; Berti, Larissa; Teixeira, João Paulo (2024). Interpretability analysis of deep models for COVID-19 detection. Artificial Intelligence in Health - AIH. ISSN 3029-2387. DOI: 10.36922/aih.2992
dc.identifier.doi10.36922/aih.2992
dc.identifier.issn3041-0894
dc.identifier.issn3029-2387
dc.identifier.urihttp://hdl.handle.net/10198/36724
dc.language.isoeng
dc.peerreviewedyes
dc.publisherAccScience Publishing
dc.relation.ispartofArtificial Intelligence in Health
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subjectCoronavirus disease 2019 detection
dc.subjectVoice processing
dc.subjectGradient-weight class activation mapping
dc.titleInterpretability analysis of deep models for COVID-19 detection
dc.typejournal article
dspace.entity.typePublication
oaire.citation.issue3
oaire.citation.titleArtificial Intelligence in Health
oaire.citation.volume1
oaire.versionhttp://purl.org/coar/version/c_970fb48d4fbd8a85
person.familyNameTeixeira
person.givenNameJoão Paulo
person.identifier663194
person.identifier.ciencia-id4F15-B322-59B4
person.identifier.orcid0000-0002-6679-5702
person.identifier.ridN-6576-2013
person.identifier.scopus-author-id57069567500
relation.isAuthorOfPublication33f4af65-7ddf-46f0-8b44-a7470a8ba2bf
relation.isAuthorOfPublication.latestForDiscovery33f4af65-7ddf-46f0-8b44-a7470a8ba2bf

Ficheiros

Principais
A mostrar 1 - 1 de 1
Miniatura
Nome:
manuscript_aih01641_PUBLICADO.pdf
Tamanho:
2.37 MB
Formato:
Adobe Portable Document Format
Ver/Abrir
Licença
A mostrar 1 - 1 de 1
Miniatura indisponível
Nome:
license.txt
Tamanho:
1.75 KB
Formato:
Item-specific license agreed upon to submission
Descrição:
Ver/Abrir

Coleções

ESTiG - Artigos em Revistas Não Indexados à WoS/Scopus