Posteroanterior Chest X-ray Image Classification with a Multilayer 1D Convolutional Neural Network-Based Classifier for Cardiomegaly Level Screening

Round 1

Reviewer 1 Report

The authors provide an interesting machine learning approach to classify and screen posteroanterior chest x-ray images for cardiomegaly with high precision.

I have only some concerns about the manuscript that should be addressed:

Within the introduction the authors provide a good overview of the medical condition and diagnosis and deepen the information for CXR image analysis. There’s a lot of information provided regarding the methods of this manuscript that should be shifted to the methods section. The introduction should include some information about classification approaches used so far for cardiomegaly or similar applications and motivate the need for the presented approach.

The authors could in addition within the methods section provide a workflow figure encompassing all the different steps to give the reader a first overview of the different steps, that are well described within the methods section.

Section 3 should be separated into pure results of the manuscript and the developed approach and then a separate Section “Discussion” should be added relating the results of this method to other methods or approaches given, if there are any comparable ones. Within the results and discussion, having a deeper look on the cases the classification method failed would provide also some interesting details of its applicability. Is there any reason seen for the failure in detection (e.g. bad image quality). How would the authors proceed in this case? Are there any other drawbacks / limitations for this approach?

Please also reformat references and formulas.

Author Response

For Reviewer: #1

The authors provide an interesting machine learning approach to classify and screen posteroanterior chest x-ray images for cardiomegaly with high precision. I have only some concerns about the manuscript that should be addressed:

Response: The point-to-point responses to all the referees are shown below. 

• Within the introduction the authors provide a good overview of the medical condition and diagnosis and deepen the information for CXR image analysis. There’s a lot of information provided regarding the methods of this manuscript that should be shifted to the methods section. The introduction should include some information about classification approaches used so far for cardiomegaly or similar applications and motivate the need for the presented approach.

Response: Thank you for reminding us. Some sentences have been added in Introduction and Section 3.2., in Pages#2-#3 and #10-#11.

• The authors could in addition within the methods section provide a workflow figure encompassing all the different steps to give the reader a first overview of the different steps, that are well described within the methods section.

Response: Thank you for reminding us. Figure 4 has been added in Section 2.4., in Page#5.

• Section 3 should be separated into pure results of the manuscript and the developed approach and then a separate Section “Discussion” should be added relating the results of this method to other methods or approaches given, if there are any comparable ones. Within the results and discussion, having a deeper look on the cases the classification method failed would provide also some interesting details of its applicability.

Response: Thank you for reminding us. Correct as suggestion. Section 3.2. has been added in Pages#10 - #11.

• Is there any reason seen for the failure in detection (e.g. bad image quality). How would the authors proceed in this case? Are there any other drawbacks / limitations for this approach?

Response: Thank you for reminding us. Some sentences have been added in Section 3.2., in Page#11.

• Please also reformat references and formulas.

Response: Thank you for reminding us. Correct as suggestion.

 

 

Author Response File: Author Response.pdf

Reviewer 2 Report

In this paper, the authors have proposed the classification of posteroanterior chest X-ray images using Multilayer 1D CNN for the screening of heart disease called cardiomegaly. Overall, the article is well written but requires some major improvements. 

1) Please consider at least 1 more dataset to verify your results. You may consider JSRT or MC datasets.

2) Please mention your contributions in the bullets form after the introduction section.

3) Separate related work from the introduction section and draw the comparison table after the related work that should mention the strengths and weaknesses of the proposed method as well as previous methods. 

4) Conclusion section is very lengthy, please trim it. Also include future work along with the conclusion. 

 

Author Response

For Reviewer: #2

In this paper, the authors have proposed the classification of posteroanterior chest X-ray images using Multilayer 1D CNN for the screening of heart disease called cardiomegaly. Overall, the article is well written but requires some major improvements. 

Response: The point-to-point responses to all the referees are shown below.

1) Please consider at least 1 more dataset to verify your results. You may consider JSRT or MC datasets.

Response: Thank you for reminding us. Some sentences have been added in Introduction (Page#2) and Table 4 has also been added in Section 3.2., in Page#10.

2) Please mention your contributions in the bullets form after the introduction section.

Response: Thank you for reminding us. Some sentences have been added in Section 3.2., in Pages#10-#11.

3) Separate related work from the introduction section and draw the comparison table after the related work that should mention the strengths and weaknesses of the proposed method as well as previous methods. 

Response: Thank you for reminding us. Some sentences and Table 4 have been added in Section 3.2., in Pages#10 - #11.

4) Conclusion section is very lengthy, please trim it. Also include future work along with the conclusion. 

Response: Thank you for reminding us. Correct as suggestion.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Most of my comments are addressed. I recommend acceptance of the article in its present form.