Challenges of deep learning methods for COVID-19 detection using public datasets
This article has been Reviewed by the following groups
Listed in
- Evaluated articles (ScreenIT)
Abstract
No abstract available
Article activity feed
-
-
SciScore for 10.1101/2020.11.07.20227504: (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 models were implemented using the Python programming language and Keras framework [51] and the experiments were carried out on a machine running Windows-10 operating system with the following hardware configuration: Intel® Core™ Pythonsuggested: (IPython, RRID:SCR_001658)Results from OddPub: Thank you for sharing your code and data.
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Although methods like GradCAM attempts to localize the most important features leading to a particular class prediction, these methods have their …
SciScore for 10.1101/2020.11.07.20227504: (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 models were implemented using the Python programming language and Keras framework [51] and the experiments were carried out on a machine running Windows-10 operating system with the following hardware configuration: Intel® Core™ Pythonsuggested: (IPython, RRID:SCR_001658)Results from OddPub: Thank you for sharing your code and data.
Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:Although methods like GradCAM attempts to localize the most important features leading to a particular class prediction, these methods have their limitations and cannot be entirely relied upon to draw conclusive analysis [39, 40]. The experiments with 2, 3, and 4-class classification tasks show that it is quite difficult for the model to distinguish between bacterial pneumonia and other viral pneumonia. Although the results, in Table 4 for CXR-Multiple-CL4, suggest that COVID Pneumonia is well distinguished from other Pneumonia, the underlying reason is very likely that the COVID Pneumonia cases come from separate data source than the other Pneumonia. Thus, we believe that further experiments with a proper dataset are required to evaluate the model’s ability to distinguish different types of Pneumonia. We suggest that in order to evaluate the model’s ability to distinguish different classes properly, it is essential to have images for each class coming from the same settings, such as the same imaging protocol, machines, demography, etc. Images from multiple settings should also be included when the objective is to assess the algorithm’s ability to work on diverse settings. However, in this case, it is essential to include images from all these settings to each of the classes to reduce the bias originating from individual settings’ peculiarities. We observed that the performance of classification results in CT-Independent-CL2 was less than CXR-Multiple-CL2. In all the experime...
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.
-
