A web-based Diagnostic Tool for COVID-19 Using Machine Learning on Chest Radiographs (CXR)
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (ScreenIT)
Abstract
This paper reports the development and web deployment of an inference model for Coronavirus COVID-19 using machine vision on chest radiographs (CXR). The transfer learning from the Residual Network (RESNET-50) was leveraged for model development on CXR images from healthy individuals, bacterial and viral pneumonia, and COVID-19 positives patients. The performance metrics showed an accuracy of 99%, a recall valued of 99.8%, a precision of 99% and an F1 score of 99.8% for COVID-19 inference. The model was further successfully validated on CXR images from an independent repository. The implemented model was deployed with a web graphical user interface for inference ( https://medics-inference.onrender.com ) for the medical research community; an associated cron job is scheduled to continue the learning process when novel and validated information becomes available.
Article activity feed
-
SciScore for 10.1101/2020.04.21.20063263: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
NIH rigor criteria are not applicable to paper type.Table 2: Resources
Software and Algorithms Sentences Resources The model was developed using python language and trained on the subset of training data with differential learning rate (10−4 - 10−2) for 15 epochs on Tesla K80 GPU from google cloud resources. pythonsuggested: (IPython, RRID:SCR_001658)Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.Results …
SciScore for 10.1101/2020.04.21.20063263: (What is this?)
Please note, not all rigor criteria are appropriate for all manuscripts.
Table 1: Rigor
NIH rigor criteria are not applicable to paper type.Table 2: Resources
Software and Algorithms Sentences Resources The model was developed using python language and trained on the subset of training data with differential learning rate (10−4 - 10−2) for 15 epochs on Tesla K80 GPU from google cloud resources. pythonsuggested: (IPython, RRID:SCR_001658)Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).
Results from LimitationRecognizer: An explicit section about the limitations of the techniques employed in this study was not found. We encourage authors to address study limitations.Results from TrialIdentifier: No clinical trial numbers were referenced.
Results from Barzooka: We did not find any issues relating to the usage of bar graphs.
Results from JetFighter: We did not find any issues relating to colormaps.
Results from rtransparent:- Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
- Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
- No protocol registration statement was detected.
-
