Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms and Hard Example Mining

Round 1

Reviewer 1 Report

Please find the detailed comments as follows:

1.       The major limitation of this work is the old and limited literature review. The literature review is not able to convey the current research status and targeted objectives based on the limitations of those works. The literature is looking around 2 years old. Very few papers from the last two years.

2.       Authors should discuss the challenging cases in the dataset and show how the proposed model is able to counter those challenges in the images.

3.       The authors should include more performance metrics in the experimental results section. In its current state, the results section is not up to the mark. It needs to be more rigorous.

4.       Authors should give proper details about the hyperparameters and quantitative details of the training model, which can be useful to ensure the reproducibility of the obtained outcomes.

5.       Authors should include some obtained visual results on some images for the proposed model and other comparative models. Please check.

6.       Please highlight the limitations of the proposed work and possible future works based on those, preferably in the conclusion section.

7.       Please check the typos mistakes and grammatical errors in the manuscript.

8.       The authors should mention the computational complexity of the models analyzed during this study.

9.       Figure 9: In different blocks, please clarify what is the last term in each block. Please explain. There are different terms in numerical terms, but what they represent should be clearer.

10.   Figure 16: Only a few joint points are shown. Figure 10 mentions different joint points (17 joint points), but not shown in Figure 16. Is Figure 16 in the same context as Figure 10?

11.   Paper organization is missing at the end of the Introduction section. Please include.

Only minor editing is required

Author Response

Dear Reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers’ constructive comments concerning our article entitled “Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Example Mining (Original title: Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Case Mining)” (ID: sustainability-2534361). These comments are all valuable and helpful for improving our article. According to your comments, we have made extensive modifications to our manuscript to make our results convincing. In revised version, changes to our manuscript were all highlighted within the document by using red-colored text. Point-by-point responses to you are listed below this letter.

Author Response File: Author Response.docx

Reviewer 2 Report

The title " Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Case Mining" is appropriate for the journal; however, the paper requires some revisions before it can be considered for publication.

-The contributions presented in the introduction section should be shortened.

-Figure 3 captions are not legible and need to be redrawn.

-More comparative studies should be added to Table 1.

-Adding a discussion section to the study will enrich it.

-Ensure that the full form of each abbreviation is provided. For example, what does CPM stand for in Line 24?

-The results of the proposed method are not presented in either the abstract or the conclusion.

-Additionally, sharing the source code on GitHub and including its link in the paper is essential for the transparency of the results.

Author Response

Dear Reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers’ constructive comments concerning our article entitled “Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Example Mining (Original title: Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Case Mining)” (ID: sustainability-2534361). These comments are all valuable and helpful for improving our article. According to your comments, we have made extensive modifications to our manuscript to make our results convincing. In revised version, changes to our manuscript were all highlighted within the document by using red-colored text. Point-by-point responses to you are listed below this letter.

Author Response File: Author Response.pdf

Reviewer 3 Report

1. Minor English needs a check. i.e. lower or Cap

2. In sec 4.1, please validate OR proof model para init.

3. The most confused part is from the last para in Sec 4.2 "Fig 18 shows ..........."   I've lost totally. Please provide numbers instead of descriptions.

4. I challenge authors to test proposal with images of disabilities. Or add an additional sec: threats to validity.

5. Finally, in the para of "In general, ...." of sec 4,2, authors mentioned that "It also improves the prediction performance of large human joint points". What is "large human joint points"? Kids?Adults?

Minor English needs to be improved.

Author Response

Dear Reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers’ constructive comments concerning our article entitled “Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Example Mining (Original title: Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Case Mining)” (ID: sustainability-2534361). These comments are all valuable and helpful for improving our article. According to your comments, we have made extensive modifications to our manuscript to make our results convincing. In revised version, changes to our manuscript were all highlighted within the document by using red-colored text. Point-by-point responses to you are listed below this letter.

Author Response File: Author Response.pdf

Reviewer 4 Report

It's a useful domain and I liked the overall work. However, it is a long article in comparison to what the idea is. Please make it succinct and then re-submit.

It is a long article in comparison to what the idea is. Please make it succinct and then re-submit. There are several grammatical and typing mistakes.

Like the following sentence: "The top-down method involves using a target detection 28 algorithm is used to identify individuals within an image,"

In some other places, the first letter of sentences was not capitalized.

There are several such errors and many sentences can be made succinct.

Author Response

Dear Reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers’ constructive comments concerning our article entitled “Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Example Mining (Original title: Improved Multi-Person 2D Human Pose Estimation Using Attention Mechanisms And Hard Case Mining)” (ID: sustainability-2534361). These comments are all valuable and helpful for improving our article. According to your comments, we have made extensive modifications to our manuscript to make our results convincing. In revised version, changes to our manuscript were all highlighted within the document by using red-colored text. Point-by-point responses to you are listed below this letter.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

necessary corrections have been made. Thank you to the authors.