Interpretability analysis of deep models for COVID-19 detection

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&rsquo;s decision-making process, including spectrograms, fundamental frequency (F0), F0 standard deviation, sex, and age. Subsequently, we examine the model&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.

Palavras-chave

Coronavirus disease 2019 detection Voice processing Gradient-weight class activation mapping

URI

http://hdl.handle.net/10198/36724

Citação

Silva; 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

Editora

AccScience Publishing

DOI

10.36922/aih.2992

Coleções

ESTiG - Artigos em Revistas Não Indexados à WoS/Scopus

Licença CC

cclicense-by

Métricas Alternativas

Ver registo completo