An Automatic Deep Learning Bowhead Whale Whistle Recognizing Method Based on Adaptive SWT: Applying to the Beaufort Sea

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. Fig1, please show your x and y label.

2. In Table1 2 kinds of "Delta f" were present, you can explain all the title in each column.

3. what means "Freq (Hz)" in Table2， central frequency or others？

4.  In Fig3， you present a pre-processing method, the data length of N  = 2500(0.25
s) and using SWT , but In Fig4， you show us a pcolor SWT Spectrum in Step2. please show me how much "N" was used in the pcolor.

5. In my opinion, you should introduce your SWT method firstly. Secondly, you can show your CNN method. Finial, the processing method and the result of your recognizing method would be indicated .

Comments on the Quality of English Language

Minor editing of English language required

Author Response

Dear reviewer, thank you for the time and effort spent reviewing the manuscript. We have made changes according to your comments, please refer to the attached document for details.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

It would greatly improve the quality of your research if you provide a more detailed and comprehensive explanation of the methodology used in the study. Specifically, please include information about the dataset used, its size, and its source. Additionally, elaborate on the criteria and rationale for choosing Rényi entropy parameters. A clear and transparent methodology will help readers understand and replicate your work.

 

To strengthen the robustness of your proposed method, consider conducting sensitivity analyses. This could involve testing your approach under different environmental conditions or with various subsets of the dataset. Providing insights into how your method performs under varying circumstances will enhance the credibility of your findings.

 

It's important to validate the performance and generalizability of your proposed method. Cross-validation is a valuable tool for assessing how well your model will perform on new, unseen data. Ensure that your conclusions are not based solely on the dataset used but can be applied to a broader context.

 

Acknowledge and discuss potential limitations and uncertainties in your research. This includes addressing any biases that might exist in the dataset and the potential impact of these biases on your results. Transparently discussing limitations demonstrates a commitment to scientific rigor.

 

In the full paper, consider incorporating statistical tests or metrics to quantify the improvements in feature detection and recognition accuracy. Statistical analysis will provide a more robust foundation for your claims.

 

Ensure that the full paper contains all the necessary details that are missing from the abstract, such as a more comprehensive methodology section, results, and discussion. The abstract provides an overview, but the full paper should offer a complete and in-depth analysis.

 

While the abstract mentions the importance of the research, make sure to provide a comprehensive list of references in the full paper that situates your work within the existing literature. This will help readers understand the broader context of your research and its contribution to the field.

 

Consider seeking peer review and collaboration with experts in the field, especially those with experience in marine biology, whale vocalization research, and deep learning. External input can help refine your methodology and ensure the highest quality of research.

 

If possible, consider making your dataset and code available to the research community. Open access to data and code promotes transparency and allows others to verify and build upon your work.

Comments on the Quality of English Language

ok

Author Response

Dear reviewer, thank you for the time and effort spent reviewing the manuscript. We have made changes according to your comments, please refer to the attached document for details.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

(1) Abstract needs to be more explicit and concise, a brief result of this research needs to be mention clearly and implication of the study should be added at the end of the abstract.

(2)    Introduction section do not meet the standard of a scientific article, authors need to revise slightly to be able to express their idea more explicit way. Please check the keywords and follow sections accordingly. On top of that Citation is required for line 30-34. Please shorten section lines 52-96 and discuss existing researchers critically and subsequently address the gap so that authors would be able to fit their research hypothesis more efficiently. Also revise lines 97-116 and provide an overview of objectives for the article.

(3)    Please follow a sequence in materials and methodology section as data, analytical techniques etc; too much of theoretical stuffs on machine learning which is indeed noisy for general readers, please change the study area map and come up with a more informative one, a index map inside the study area map is highly demanding for the global audience. 

(4)    I do not see a proper discussion section in your article, authors might come up with a good bunch of results and a critical discussion with cross referencing is indispensable her. I would rather suggest to authors to provide two sections one is result and other is a comprehensive discussion. Result section does carry a lot of information which need to be shorten. I also do not see limitation and future direction in discussion section. Please revise conclusion and put limitation and future direction in discussion section.  

 

Author Response

Dear reviewer, thank you for the time and effort spent reviewing the manuscript. We have made changes according to your comments, please refer to the attached document for details.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The objectives of the paper should be included in the abstract of the paper.

There are similarities between the abstract and the introduction. Rewrite the abstract

Section 6 should be modified to "6. Discussion Conclusion"

The author should focus on the mathematical analysis in a section

Comments on the Quality of English Language

Minor

Author Response

Dear reviewer, thank you for the time and effort spent reviewing the manuscript. We have made changes according to your comments, please refer to the attached document for details.

Author Response File: Author Response.pdf

Reviewer 5 Report

Comments and Suggestions for Authors

In this paper, the recognition of whale signals is done well by using intelligent learning method and the method is innovative. Meets the journal's requirements for a research article and is recommended for acceptance with minor revisions.However, the content of this article still has some problems, as follows.

1.Content of research

1)The abstract of this article should be further shortened and should have a brief result on the application of the algorithm.

2)From the line 358, if possible there should be some comparisons with published articles here, not only with CNN or LSTM.

 

Comments on the Quality of English Language

2.English Writing

1)In line 10, 106, 110 ,184, these senteces are a bit long, not express the meaning clearly. If possible turn into short one.

2)Such as in line 12, 1, 24, 44, there are incorrect grammar.  If possible, modify them.

3)In line 24, there should 79.54%, you missed "%".

4)From line 242 to 244 , there is repetition of preceding and following expressions.

Author Response

Dear reviewer, thank you for the time and effort spent reviewing the manuscript. We have made changes according to your comments, please refer to the attached document for details.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

1. The introduction provides a clear overview of the importance of bowhead whales and the need for accurate recognition techniques. However, it lacks specific background information about the current state of bowhead whale recognition and the challenges involved. Including this context would make the research more informative.

2. The use of adaptive Stationary Wavelet Transform (SWT) for feature extraction is a novel approach, but the paper doesn't provide enough details about how this technique is applied. It would be helpful to include a brief explanation of how adaptive SWT works and why it was chosen over other methods.

3. The paper mentions that the adaptive SWT outperforms Short-Time Fourier Transform (STFT) in terms of Signal-to-Clutter Ratio (SCR), but it doesn't provide any comparative analysis or statistical tests to support this claim. Including a detailed comparison or statistical results would strengthen the paper.

4. The CNN-LSTM deep learning model is introduced, but there is a lack of information about its architecture and how it is specifically used for bowhead whale voice recognition. Readers would benefit from more details about the model's structure and training process.

5. The reported recognition accuracy of 92.85% is impressive, but it would be helpful to know the size of the dataset and the number of samples used for training and testing to understand the statistical significance of this result.

6. The paper mentions consistency between the recognition results in the Beaufort Sea and results from a fisheries ecology study, but it doesn't provide any specific details or data from these comparisons. Including this data or at least a brief explanation of how the comparisons were made would strengthen the paper's argument.

7. The conclusion mentions the "great significance" of the results but does not elaborate on the practical implications or potential applications of the research. Adding a discussion of how this work can be used for bowhead whale conservation and ocean acoustic remote sensing technology would be beneficial.

8. The language and writing style in the summary need improvement. Some sentences are quite long and complex, making it challenging for readers to follow the content smoothly. Simplifying and breaking down complex ideas into smaller sentences would enhance readability.

9. Citations and references are missing from the summary. It is important to provide proper credit to previous work and research that have influenced or supported the current study. Adding citations where necessary is essential.

10. Proofreading is needed for grammar and punctuation. There are a few grammatical errors and awkward phrasings that should be addressed to improve the overall quality of the summary.

Comments on the Quality of English Language

ok

Author Response

Dear reviewer, thank you again for your time and effort spent reviewing the manuscript. It is due to your insightful comments that our paper can be continuously improved. We have made changes based on  your comments, please see the attached document for details. Thank you again!

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I'm happy with the revision

Author Response

Dear reviewer, thank you again for your time and effort spent reviewing the manuscript. Regarding the English grammar and writing,  we have consulted English writing experts and have revised and proofread the revised manuscript. please see the revisied manuscript for details. Thank you again!