Next Article in Journal
Experimental Study on Smoke Spreading Laws for Tunnel Fires with Moving Fire Source
Previous Article in Journal
Numerical Simulation of Flow and Flame Dynamics of a Pool Fire Under Combined Effects of Wind and Slope
 
 
Article
Peer-Review Record

Optimizing Fire Scene Analysis: Hybrid Convolutional Neural Network Model Leveraging Multiscale Feature and Attention Mechanisms

by Shakhnoza Muksimova 1, Sabina Umirzakova 1,*, Mirjamol Abdullaev 2 and Young-Im Cho 1,*
Reviewer 1:
Reviewer 2:
Submission received: 21 October 2024 / Revised: 12 November 2024 / Accepted: 17 November 2024 / Published: 20 November 2024

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. Original Submission

1.1. Recommendation

Major Revision

Title: Optimizing Fire Scene Analysis: Hybrid Convolutional Neural Network Model Leveraging Multiscale Features and Attention Mechanisms

Overview and General Recommendations:

Fire is a threat to human property and safety, and early prevention and prediction are important issues in fire prevention. In particular, with the development of AI, various learning algorithms (AlexNet, Inception, VGG, Residual Network) are being used to detect the initial smoke, which is the cause of fire, and to improve the early fire detection capability and accuracy. This paper introduces a new hybrid Convolutional Neural Network Model to improve fire detection capability and to develop it for field deployment, and has high academic and research value.

This paper consists of an introduction, review, method, results, discussion, and conclusion, and summarizes various parts in detail. As a result, this paper is judged to be suitable for publication in Fire, but major revision of the content is necessary.

The detailed main comments of the paper are as follows.

1. Introduction

The introduction section and the review section need to be combined.

- Since the introduction section should include an introduction, review, and research focus, it is recommended to write them together rather than inserting the review section separately.

2. Method

- The method section introduces various models used in the study and provides detailed information on how to use them.

3. Experimental setup and results

- The experimental results are produced by applying 3,000 data for each fire, so the results are credible.

- It is a good idea to compare the results with other models, but it is necessary to summarize only the results for the developed model.

- It is considered a good idea to summarize all the results of the developed model in the results section and include the results of the comparison in the discussion section.

4. Discussion

- It is recommended that the comparison with the existing model written in the results section be included in the discussion section to summarize the superiority and limitations of the development technology.

5. Conclusion

The entire paper is clearly summarized well.

7. References

Most references are necessary to change to the MDPI format.

Author Response

We sincerely thank the reviewers for their thorough and constructive feedback. Your comments have greatly contributed to improving the clarity, depth, and overall quality of our manuscript. We appreciate your insights and suggestions, which have enabled us to strengthen our work and provide a more comprehensive presentation of our research. Thank you for your time and effort in reviewing our paper. We made our reply on the file below: 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In general, a very well-written and descriptive manuscript. My only concern is that the Discussion section is completely inadequate. The paper does a very good job of describing the field, the development of the approach and model and its application. It falls significantly short in assessing the importance of their findings to the model construction, the importance of their accomplishments to fire modeling specifically, and the role or significance of their work to the overall field of modelling using AI approaches. Without this discussion, the manuscript is not publishable.

Minor Edits:

Line 117: Delete "authors in"

Line 120: Delete "The authors of"

Line 125: Delete "The authors in"

Line 128: Delete "Reference"

Line 135: Delete "The authors in"

Line 138: Delete "Reference"

Line 143: Delete "Reference"

Line 145: Delete "Reference"

Line 170: Add [25] after ResNet 50

Line 193: Add [26] after VCG16

Line 211: Add [27] after EfficientNet-B3

Table 1: Amount of Data is not descriptive enough - 3000 what

 

Author Response

We sincerely thank the reviewers for their thorough and constructive feedback. Your comments have greatly contributed to improving the clarity, depth, and overall quality of our manuscript. We appreciate your insights and suggestions, which have enabled us to strengthen our work and provide a more comprehensive presentation of our research. Thank you for your time and effort in reviewing our paper. We made our reply on the file below: 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The author has corrected most of the points pointed out by the reviewer. Therefore, this paper is suitable for publication in Fire.

However, the MDPI format of the reference needs to be checked.

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed my concerns as stated in the initial review.

Back to TopTop