.png)
Artificial Intelligence-Driven Eye Disease Classification Model
Round 1
Reviewer 1 Report
1. In the abstract, provide variations in sentences, most sentences starts with "The", this affects the flow of reading as it is repetitive.
2. Although this is just a draft, consider following the MDPI Journal template on the next submission.
3. In the review of literature make it thematic, suggestion will be a paragraph for traditional machine learning on the 1st section, the succeeding section is on deep learning. Do not mix the two on the literature. It makes the reader more confused.
4. State already on line 92, 93 what kind of feature extraction used, what feature selection technique is used that previous studies did not apply.
5. Figure 2 needs to be improved, it lacks details on the pipeline or the framework, add more details on the process. Plus figures are encouraged to have DPI of 600 or above so that it will not blur.
6. Equations are not properly align.
7. Again Figure 3 can be improve further, it is too plain and simple for a complex operation.
8. Figure 4 and Figure 5 is stretched, at least make it presentable.
9. At least provide brief explanation on the evaluation metrics, how it is measured, include formula.
10. Where is the hyperparameter optimization section? Provide one.
11. The paper lacks a clear explanation of the specific limitations encountered during the classification of fundus images, hindering a comprehensive understanding of the challenges faced. Additionally, more details are needed regarding the experimental setup, such as hyperparameter settings, model architecture specifications, and the rationale for the choice of specific optimization algorithms.
12. The paper's results are promising, but they lack comprehensive reporting. To strengthen the conclusions, the paper should include detailed performance metrics and statistical analysis to support the claims made. Furthermore, the conclusions could be improved by discussing the practical implications and potential real-world applications of the proposed EDC model.
Sentences lacks variations, it is hard to read. Grammar need major revisions.
Author Response
Dear Editor and Reviewers,
I thank Editor and Reviewers for their suggestions in improving the standard of the article. I addressed the reviewers’ comments and modified the article.
Reviewer 1:
- In the abstract, provide variations in sentences, most sentences starts with "The", this affects the flow of reading as it is repetitive.
Response: I thank you for your valuable suggestion. I updated the Abstract as per the suggestion.
- Although this is just a draft, consider following the MDPI Journal template on the next submission.
Response: As per the suggestion, I updated the manuscript in MDPI template.
- In the review of literature make it thematic, suggestion will be a paragraph for traditional machine learning on the 1st section, the succeeding section is on deep learning. Do not mix the two on the literature. It makes the reader more confused.
Response: As per the suggestion, section 2: Literature review is introduced in line no. 106.
- State already on line 92, 93 what kind of feature extraction used, what feature selection technique is used that previous studies did not apply.
Response: As per the comment, I introduced the advantages of feature extraction and selection technique in line numbers 87 and 93.
- Figure 2 needs to be improved, it lacks details on the pipeline or the framework, add more details on the process. Plus figures are encouraged to have DPI of 600 or above so that it will not blur.
Response: As per the suggestion, Figure 1 presents the framework and I improved the quality of the figure 2. Additionally, all figures are changed to 800 DPI
- Equations are not properly align.
Response: As per the comment, the equations are aligned in the updated manuscript.
- Again Figure 3 can be improve further, it is too plain and simple for a complex operation.
Response: As per the suggestion, I included additional details into the figure 3.
- Figure 4 and Figure 5 is stretched, at least make it presentable.
Response: I thank you for the suggestion, I changed both figures 4 and 5.
- At least provide brief explanation on the evaluation metrics, how it is measured, include formula.
Response: I included the relevant metrics from line no. 351.
- Where is the hyperparameter optimization section? Provide one.
Response: As per the suggestion, I introduced a sub section 3.6 for hyperparameter in line number 327.
- The paper lacks a clear explanation of the specific limitations encountered during the classification of fundus images, hindering a comprehensive understanding of the challenges faced. Additionally, more details are needed regarding the experimental setup, such as hyperparameter settings, model architecture specifications, and the rationale for the choice of specific optimization algorithms.
Response: I included the specific limitations in section 5, line number 528. And discussed the hyperparameter settings and model architecture in section 4, line number 365 and Table 2.
- The paper's results are promising, but they lack comprehensive reporting. To strengthen the conclusions, the paper should include detailed performance metrics and statistical analysis to support the claims made. Furthermore, the conclusions could be improved by discussing the practical implications and potential real-world applications of the proposed EDC model.
Response: As per the comment, I introduced a detailed explanation in line number 455 in the section 5. In addition, I included the practical implications in the section 6 conclusion part, line number 563.
- Grammatical issues
Response: I requested Dr. Aruna Devi, Cambridge Institute of Technology, India to proofread the article.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The authors propose a DL-based EDC to identify eye diseases from complex retinal images. The paper is well-written and concise. I have some suggestions and concerns for the authors:
1. Many acronyms have been used throughout the paper without initial explanation. Consider providing brief explanations upon the first mention for clarity.
2. It would be beneficial to explicitly state the research problem or objective to provide a clearer context for readers. Elaborate a bit more on the motivation for the research.
3. What about sensitivity analysis of the proposed models?? What are your suggestions and ideas about this task?
4. The conclusion appears somewhat unclear. It is advisable for the authors to provide quantifiable explanations demonstrating why their proposed model outperforms others.
5. Limited suggestions for future work in the last section and underwhelm, need to be revised, these need to be clearly exploited.
There are English grammar errors in the text.
Author Response
Dear Editor and Reviewers,
I thank Editor and Reviewers for their suggestions in improving the standard of the article. I addressed the reviewers’ comments and modified the article.
Reviewer 2:
The authors propose a DL-based EDC to identify eye diseases from complex retinal images. The paper is well-written and concise. I have some suggestions and concerns for the authors:
- Many acronyms have been used throughout the paper without initial explanation. Consider providing brief explanations upon the first mention for clarity.
Response: As per the suggestion, I have added the explanation for the acronym at the first time in the content.
- It would be beneficial to explicitly state the research problem or objective to provide a clearer context for readers. Elaborate a bit more on the motivation for the research.
Response: As per the comment, I have elaborated the motivation in line number 76.
- What about sensitivity analysis of the proposed models?? What are your suggestions and ideas about this task?
Response: I thank you for the suggestion, I have included both sensitivity and specificity in the results.
- The conclusion appears somewhat unclear. It is advisable for the authors to provide quantifiable explanations demonstrating why their proposed model outperforms others.
Response: I have modified the conclusion part (line number 556) as per the suggestion.
- Limited suggestions for future work in the last section and underwhelm, need to be revised, these need to be clearly exploited.
Response: As per the suggestion, I included the future work in section 5, line number 540.
- There are English grammar errors in the text.
Response: I requested Dr. Aruna Devi, Cambridge Institute of Technology, India to proofread the article.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
1. There are blank spaces not utilized on some pages. Just a minor fix.
2. Major improvements are noted.
