The Practicality of Deep Learning Algorithms in COVID-19 Detection: Application to Chest X-ray Images

Round 1

Reviewer 1 Report

This paper attempts to assess the feasibility of using convolutional neural networks for analysis of pulmonary radiography images distinguishing COVID-19 infections from non-infected cases and from other type of viral or bacterial pulmonary affections. In summary, the paper is organized well. I appreciate the motivation and novelty of the proposed study. However, it should be revised majorly before it can be considered to be published.

1. In the abstract, it is expected to present more results of the performance.
2. The background of the proposed study should be further explained in detail. Some concepts are hard to comprehend without explaining clearly.
3. Please explore the robustness of the proposed method.
4. Spiking neural network is biologically inspired and low-power-consumption. Please discuss the comparison between the proposed model and other SNN models, including: Efficient Spike-Driven Learning With Dendritic Event-Based Processing; Spatial properties of STDP in a self-learning spiking neural network enable controlling a mobile robot.
5. Please further add more experiments on the comparison with more complicated data set.
6. Please discuss the application of the proposed study in other digital neuromorphic computing, including: CerebelluMorphic; BiCoSS; Scalable digital neuromorphic architecture for large-scale biophysically meaningful neural network with multi-compartment neurons; Real-time neuromorphic system for large-scale conductance-based spiking neural networks.
7. Can you compare more aspects of the proposed study with the state-of-the-art works. Detailed comparison is vital and meaningful to illustrate the novelty and advantages of the proposed work.
8. Grammar should be further improved.

Author Response

I would like to thank the anonymous reviewer for spending his/her valuable time reviewing and editing my paper. His/her feedback is very helpful and constructive. I am really happy that the reviewer appreciates the novelty of the paper, and considers the paper is well organized. I am open to more revisions and edits (if required) until the reviewer’s ultimate satisfaction is reached. My responses to each concern raised by the reviewer can be found in the attached document.

Author Response File: Author Response.pdf

Reviewer 2 Report

The manuscript is well-written, and the methodology is clearly described. I had no problem going through it. Please consider my feedback below.

(1)

My view is that the study is not introducing a novel method nor a new area of application. Given the plethora of works published in the COVID context, I am not sure if such study would be a considerable addition in favour of Deep Learning in this regard.

Perhaps the study may emphasise further on explainability, for example. This might help introduce further aspects compared to the literature, using the same models with very close performance.

(2)

I recommend evaluating the models based on cross-validation. The dataset used is quite small, and cross-validation should have been more suitable for evaluation.

(3)

For clarity, I recommend presenting the related work in a separate section aside from the introduction.

(4)

Please use the reference below for TensorFlow.

Abadi, M.; Barham, P.; Chen, J.; Chen, Z.; Davis, A.; Dean, J.; Devin, M.; Ghemawat, S.; Irving, G.; Isard, M.; & et al. Tensorflow: A system for large-scale machine learning. In Proceedings of the 12th (USENIX) Symposium on Operating Systems Design and Implementation (OSDI 16), 2–4 November 2016. Savannah, GA, USA; pp. 265–283.

Author Response

I would like to thank the anonymous reviewer for spending his/her valuable time reviewing and editing my paper. His/her feedback is very helpful and constructive. I am really happy that the reviewer considers the paper is well-written, and the proposed methodology is clearly described. I am open to more revisions and edits (if required) until the reviewer’s ultimate satisfaction is reached. My responses to each concern raised by the reviewer can be found in the attached document.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Thanks very much for accommodating the feedback.
One last point, please elaborate further on the explainability part at the abstract and conclusions.