Potato Leaf Disease Segmentation Method Based on Improved UNet

Round 1

Reviewer 1 Report

The study is interesting but needs some improvements.

 

The abstract needs to present the study problem more clearly with a clear hypothesis. This hypothesis should also be included in the introduction. The methods need to be better described, it is very confusing when reading. It is important to try to present a chronological order. Furthermore, figure 1 is of poor quality, which makes interpretation difficult. Figures 6 and 7 require better quality as well as the placement of some type of scale to identify the change in color as well as orientation. Figure 10 also needs better quality and color scale or some type of quantification by area in the various photos for a better interpretation of the results.

 

Author Response

Dear Professor,

Thank you very much for your advice, Here is our response to the suggestions you raised:

1. We noticed in your response that there could be a clearer presentation of the issues in the abstract and introduction. Therefore, we have made modifications to the content in both the abstract and introduction to better clarify the problems.
2. In the methodology section, we have restructured it chronologically, starting with the basic framework of UNet, followed by introductions to the models used, and concluding with the improved model.
3. For Figures 1, 6, 7, and 10, we have replaced them with higher-quality images for presentation.

Thank you very much for your suggestions. Your feedback was very insightful, and the paper has been revised and responded to in detail. In this response , "applsci-2606967 Revised.pdf" refers to a reply regarding the revised version of the entire article.

Author Response File: Author Response.pdf

Reviewer 2 Report

Please add clear objectives in the last paragraph of the introduction

Please explain how leaves were treated for disease inoculation in method section

Please add (https://doi.org/10.47278/journal.ijab/2023.059; https://doi.org/10.47278/journal.ijab/2023.049) to improve introduction and discussion section

Please add a short future perspective before conclusions how this improved Unet is more reliable than other methods

 

 

Moderate English revisions 

Author Response

Dear Professor,

Thank you very much for your advice, Here is our response to the suggestions you raised:

1. Following your suggestion, a clear objective has been added in the final paragraph of the introduction section.
2. In section 2.1 of the methodology, details on how the leaves were processed have also been added.
3. Discussion sections have been added in both section 2 (Methodology) and section 4.
4. Brief future perspectives have been added in the discussion section just before section 5 (Conclusion).                                                                        

Thank you very much for your suggestions. Your feedback was very insightful, and the paper has been revised and responded to in detail. In this response , "applsci-2606967 Revised.pdf" refers to a reply regarding the revised version of the entire article.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this work, Fu et al. propose an innovative approach, RS-UNet, which combines ResNet50 with UNet to address information loss and accuracy issues in potato leaf disease image segmentation. Their method incorporates SE attention mechanisms, achieving a 6.33% Dice coefficient improvement, surpassing other networks, and offering practical advantages in training time and disease spot segmentation accuracy for potato leaf disease analysis.

It is an interesting study. Please find my minor comments:

1) It would be beneficial to elaborate on the implementation of the SE attention mechanism and how it contributes to feature selection and improved accuracy.

2) Could you provide more details on the specific potato characteristics integrated into the UNet network to enhance its performance and relevance to potato leaf disease?

3) At the end of Introduction, please add a line on novelty of the current work.

Minor editing of English language required

Author Response

Dear Professor,

Thank you very much for your advice, Here is our response to the suggestions you raised:

1. Regarding the questions you raised, I have provided a detailed explanation of the implementation of the SE attention mechanism in section 2.2.3 and how it enhances feature extraction and accuracy.
2. In section 2.2 of the methodology, I have provided a more detailed explanation of how the network improves performance, with specific additions to section 2.2.2 (Residual Modules), section 2.2.3 (SE Attention Mechanism), and section 2.3 (Model Enhancement).
3. At the end of the introduction in section 1, I have added information highlighting the novelty of this work.

Thank you very much for your suggestions. Your feedback was very insightful, and the paper has been revised and responded to in detail. In this response , "applsci-2606967 Revised.pdf" refers to a reply regarding the revised version of the entire article.

Author Response File: Author Response.pdf

Reviewer 4 Report

The abstract is confusing because from the beginning it talks about results. It should be structured as follows: introduction, objective, main results and conclusion.

In materials and methods there is a missing section on what type of statistics was performed. What program was used and what graphs or data will be used to analyze the results.

The discussion is one of the most important parts of a scientific article. Here the results should be discussed in comparison with other research. More paragraphs should be added in the discussion. Possible limitations should also be added.

It is necessary to indicate if there is financing to know if there is a possible conflict of interest.

Moderate editing of English language required

Author Response

Dear Professor,

Thank you very much for your advice, Here is our response to the suggestions you raised:

1. In response to the issues you raised, modifications have been made in the abstract section.
2. In the Materials and Methods section, details regarding the type used are provided in section 2.1, and the procedures and environment used are described in section 3.1.
3. A discussion section has been added in part 4 for comparative analysis.
4. At the end of the article, it is stated that this project had no funding and no conflicts of interest exist.

Thank you very much for your suggestions. Your feedback was very insightful, and the paper has been revised and responded to in detail. In this response , "applsci-2606967 Revised.pdf" refers to a reply regarding the revised version of the entire article.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I Don´t have more questions. I suggest accepting.

Author Response

Dear Professor,

We sincerely express our gratitude for your professor's review, comments, and suggestions on our submitted article. Thanks to your expert and thorough examination and feedback, we were able to conduct an in-depth study of the article, leading to its improvement and refinement.

Sincerely.

Fu Jun.

Reviewer 2 Report

Article is good but discussion is a weaker part of the manuscript. Authors should add suggested citations as well as related citations to improve this section. I am shocked how authors completed the discussion section without citations of related study.

Author Response

Dear Professor,

Thank you very much for your advice, Here is our response to the suggestions you raised:

The latest revisions involve modifications and enhancements to the discussion section, thereby enriching the entire article. Additionally, references supporting the entire article have been included.

Thank you very much for your suggestions. Your feedback was very insightful, and the paper has been revised and responded to in detail. In this response , "applsci-2606967 Revised.pdf" refers to a reply regarding the revised version of the entire article.

Author Response File: Author Response.pdf