TY - JOUR
T1 - Improving Uncertainty Estimation with Semi-Supervised Deep Learning for COVID-19 Detection Using Chest X-Ray Images
AU - Calderon-Ramirez, Saul
AU - Yang, Shengxiang
AU - Moemeni, Armaghan
AU - Colreavy-Donnelly, Simon
AU - Elizondo, David A.
AU - Oala, Luis
AU - Rodríguez-Capitán, Jorge
AU - Jiménez-Navarro, Manuel
AU - Lopez-Rubio, Ezequiel
AU - Molina-Cabello, Miguel A.
N1 - Publisher Copyright:
© 2013 IEEE.
PY - 2021
Y1 - 2021
N2 - In this work we implement a COVID-19 infection detection system based on chest X-ray images with uncertainty estimation. Uncertainty estimation is vital for safe usage of computer aided diagnosis tools in medical applications. Model estimations with high uncertainty should be carefully analyzed by a trained radiologist. We aim to improve uncertainty estimations using unlabelled data through the MixMatch semi-supervised framework. We test popular uncertainty estimation approaches, comprising Softmax scores, Monte-Carlo dropout and deterministic uncertainty quantification. To compare the reliability of the uncertainty estimates, we propose the usage of the Jensen-Shannon distance between the uncertainty distributions of correct and incorrect estimations. This metric is statistically relevant, unlike most previously used metrics, which often ignore the distribution of the uncertainty estimations. Our test results show a significant improvement in uncertainty estimates when using unlabelled data. The best results are obtained with the use of the Monte Carlo dropout method.
AB - In this work we implement a COVID-19 infection detection system based on chest X-ray images with uncertainty estimation. Uncertainty estimation is vital for safe usage of computer aided diagnosis tools in medical applications. Model estimations with high uncertainty should be carefully analyzed by a trained radiologist. We aim to improve uncertainty estimations using unlabelled data through the MixMatch semi-supervised framework. We test popular uncertainty estimation approaches, comprising Softmax scores, Monte-Carlo dropout and deterministic uncertainty quantification. To compare the reliability of the uncertainty estimates, we propose the usage of the Jensen-Shannon distance between the uncertainty distributions of correct and incorrect estimations. This metric is statistically relevant, unlike most previously used metrics, which often ignore the distribution of the uncertainty estimations. Our test results show a significant improvement in uncertainty estimates when using unlabelled data. The best results are obtained with the use of the Monte Carlo dropout method.
KW - chest x-ray
KW - computer aided diagnosis
KW - Coronavirus
KW - Covid-19
KW - MixMatch
KW - semi-supervised deep learning
KW - Uncertainty estimation
UR - http://www.scopus.com/inward/record.url?scp=85107340495&partnerID=8YFLogxK
U2 - 10.1109/ACCESS.2021.3085418
DO - 10.1109/ACCESS.2021.3085418
M3 - Article
AN - SCOPUS:85107340495
SN - 2169-3536
VL - 9
SP - 85442
EP - 85454
JO - IEEE Access
JF - IEEE Access
M1 - 9445026
ER -