ResNet Based on Multi-Feature Attention Mechanism for Sound Classification in Noisy Environments
Round 1
Reviewer 1 Report
This study proposes the combination of ResNet with an SE module to improve the accuracy of sound classification. The topic is timely and would be attractive to many readers. Additionally, the findings are reasonable and worthy of reporting. However, the structure of the paper and the presentation of the findings need improvement.
-
The introduction includes many technical terms, but they are not properly defined or explained. Furthermore, in Chapters 1 and 2, the ways to highlight the limitations of previous studies and the necessity of this study are poorly written. These chapters need to be rewritten for the readers' understanding.
-
There are too many unnecessary figures in the manuscript. Please consider omitting them to make the manuscript more concise.
-
In Figure 11, please provide an explanation of the boxplots, including measures such as median and quartiles.
-
Regarding Figure 11, the median of the new model is greater than the others, but it shows larger variations. It would be helpful to explain the reason behind this.
-
Table 2 is not easy to read. Is there any way to present these values in a clearer manner?
-
In Table 3, the difference between GoogleNet and this model is minimal. Please discuss which one would be better for the users and provide your insights.
There are many typos, punctuation errors (e.g., spacing), and mixed uses of different tenses in the manuscript. Therefore, it is strongly encouraged to thoroughly proofread it again.
Author Response
Dear Reviewer,
Hello! I have attached the revised manuscript along with my response to your comments. Please find them in the attached files.
Thank you for your time and valuable feedback. I look forward to hearing your thoughts on the revised version.
Best regards,
Xiaoyu Yuan.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
1. How is phonemic classification used in educational areas in different fields?
2. How can acoustic scene classification accurately detect and control the source of noise? What is the role of acoustic landscape classification for smart wearable devices and robotic navigation? Please explain in detail please.
3. The source of basic equations must be written.
4. It could be very interesting if authors can provide real materials with assumed material properties. How these material parameters were obtained? Otherwise, the problem looks like a mathematical exercise.
5. What is the main objective behind the current study? It is beneficial for the readers to add more explanations about the novel contribution of this method from theoretical/experimental viewpoints.
6. - The authors must explicitly declare the assumptions and limitations of their model. It seems its application is quite restricted while computational methods are capable of dealing with a much wider application range.
7. - The introduction part needs to be extended by discussing more relevant papers. The authors should appropriately extend this section by discussing more relevant works focusing on different methods and models in the literature. Specifically, the authors should carefully discuss previous papers.
8. - It is suggested to add a more in-depth explanation of the model, its justification, and more discussions on the results.
9. - Is the presented method capable of solving the nonlinear behavior of the system? Please explain the procedure if any.
10. - The paper should be carefully double-checked from a grammatical point of view.
11. The results presented in this study were not verified and no one can check the integrity of the mathematical modeling.
12. The obtained findings of this work should be compared to experimental results or at least with other published results in the literature.
13. The authors should try to give advantages of using their method compared to others.
14. The authors need to explain that the numerical approach used in the research is one of the appropriate solutions in the context of the research problem. What are the achievements of previous studies based on a numerical basis? Also, describe what has not been achieved.
Moderate editing of English language required
Author Response
Dear Reviewer,
Hello! I have attached the revised manuscript along with my response to your comments. Please find them in the attached files.
Thank you for your time and valuable feedback. I look forward to hearing your thoughts on the revised version.
Best regards,
Xiaoyu Yuan.
Author Response File: Author Response.pdf
Reviewer 3 Report
In my opinion, the presentation of results, discussions and conclusions should be more detailed.
Moderate editing of English language required
Author Response
Dear Reviewer,
Hello! I have attached the revised manuscript along with my response to your comments. Please find them in the attached files.
Thank you for your time and valuable feedback. I look forward to hearing your thoughts on the revised version.
Best regards,
Xiaoyu Yuan.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The revision is satisfactory.
There are still punctuation errors, and I'm not sure if they can be edited in a proofreading stage.
Reviewer 2 Report
After reviewing the revised version, it was found that:
The authors have made the requested improvements
The authors have answered many concerns
I think now the manuscript can be accepted for publication in the esteemed journal
