Improving Uncertainty Estimation with Semi-Supervised Deep Learning for COVID-19 Detection Using Chest X-Ray Images

Saul Calderon-Ramirez, Shengxiang Yang, Armaghan Moemeni, Simon Colreavy-Donnelly, David A. Elizondo, Luis Oala, Jorge Rodríguez-Capitán, Manuel Jiménez-Navarro, Ezequiel Lopez-Rubio, Miguel A. Molina-Cabello

Research output: Contribution to journalArticlepeer-review

Abstract

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.

Original languageEnglish
Article number9445026
Pages (from-to)85442-85454
Number of pages13
JournalIEEE Access
Volume9
DOIs
Publication statusPublished - 2021
Externally publishedYes

Keywords

  • chest x-ray
  • computer aided diagnosis
  • Coronavirus
  • Covid-19
  • MixMatch
  • semi-supervised deep learning
  • Uncertainty estimation

Fingerprint

Dive into the research topics of 'Improving Uncertainty Estimation with Semi-Supervised Deep Learning for COVID-19 Detection Using Chest X-Ray Images'. Together they form a unique fingerprint.

Cite this