| Author | Chagas, Paulo | |
| Author | Souza, Luiz | |
| Author | Pontes, Izabelle | |
| Author | Calumby, Rodrigo | |
| Author | Angelo, Michele | |
| Author | Duarte, Angelo | |
| Author | Santos, Washington L. C. dos | |
| Author | Oliveira, Luciano | |
| Access date | 2023-02-02T16:58:53Z | |
| Available date | 2023-02-02T16:58:53Z | |
| Document date | 2022 | |
| Citation | CHAGAS, Paulo et al. Uncertainty-aware membranous nephropathy classification: A Monte-Carlo dropout approach to detect how certain is the model. Computer Methods in Biomechanics and Biomedical Engineering, p. 1-11, 2022. | en_US |
| ISSN | 1476-8259 | en_US |
| URI | https://www.arca.fiocruz.br/handle/icict/56838 | |
| Sponsorship | Fundação de Apoio à Pesquisa do Estado da Bahia (FAPESB).
Inova Fiocruz - Ideias inovadoras.
Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq). | en_US |
| Language | eng | en_US |
| Publisher | Taylor and Francis Group | en_US |
| Rights | open access | en_US |
| Subject in Portuguese | Nefropatia membranosa | en_US |
| Subject in Portuguese | Aprendizado profundo | en_US |
| Subject in Portuguese | Estimativa de incerteza | en_US |
| Title | Uncertainty-aware membranous nephropathy classification: a Monte-Carlo dropout approach to detect how certain is the model | en_US |
| Type | Article | en_US |
| DOI | 0.1080/21681163.2022.2029573 | |
| Abstract | Membranous nephropathy (MN) is among the most common glomerular diseases that cause nephrotic syndrome in adults. To aid pathologists on performing the MN classification task, we proposed here a pipeline consisted of two steps. Firstly, we assessed four deep-learning-based architectures, namely, ResNet-18, MobileNet, DenseNet, and Wide-ResNet. To achieve more reliable predictions, we adopted and extensively evaluated a Monte-Carlo dropout approach for uncertainty estimation. Using a 10-fold cross-validation setup, all models achieved average F1-scores above 92%, where the highest average value of 93.2% was obtained by using Wide-ResNet. Regarding uncertainty estimation with Wide-ResNet, high uncertainty scores were more associated with erroneous predictions, demonstrating that our approach can assist pathologists in interpreting the predictions with high reliability. We show that uncertainty-based thresholds for decision referral can greatly improve classification performance, increas-ing the accuracy up to 96%. Finally, we investigated how the uncertainty scores relate to complexity scores defined by pathologists | en_US |
| Affilliation | Universidade Federal Da Bahia. IvisionLab. Salvador, BA, Brasil. | en_US |
| Affilliation | Universidade Federal Da Bahia. IvisionLab. Salvador, BA, Brasil. | en_US |
| Affilliation | Fundação Oswaldo Cruz. Instituto Gonçalo Moniz. Salvador, BA, Brasil. | en_US |
| Affilliation | Universidade Estadual de Feira de Santana. Departamento de Tecnologia. Salvador, BA, Brasil. | en_US |
| Affilliation | Universidade Estadual de Feira de Santana. Departamento de Tecnologia. Salvador, BA, Brasil. | en_US |
| Affilliation | Universidade Estadual de Feira de Santana. Departamento de Tecnologia. Salvador, BA, Brasil. | en_US |
| Affilliation | Fundação Oswaldo Cruz, Instituto Gonçalo Moniz, Salvador, BA, Brasil. | en_US |
| Affilliation | Universidade Federal Da Bahia. IvisionLab. Salvador, BA, Brasil. | en_US |
| Subject | Membranous nephropathy | en_US |
| Subject | Deep learning | en_US |
| Subject | Uncertainty estimation | en_US |
| DeCS | Nefropatias | en_US |
| DeCS | Aprendizado profundo | en_US |
| DeCS | Incerteza | en_US |
| xmlui.metadata.dc.subject.ods | 04 Educação de qualidade | |
