All reviews of published articles are made public. This includes manuscript files, peer review comments, author rebuttals and revised materials. Note: This was optional for articles submitted before 13 February 2023.
Peer reviewers are encouraged (but not required) to provide their names to the authors when submitting their peer review. If they agree to provide their name, then their personal profile page will reflect a public acknowledgment that they performed a review (even if the article is rejected). If the article is accepted, then reviewers who provided their name will be associated with the article itself.
Dear authors, we are pleased to verify that you meet the reviewer's valuable feedback to improve your research.
Thank you for considering PeerJ Computer Science and submitting your work.
Kind regards
PCoelho
[# PeerJ Staff Note - this decision was reviewed and approved by Shawn Gomez, a PeerJ Section Editor covering this Section #]
**PeerJ Staff Note:** Please ensure that all review, editorial, and staff comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** When you prepare your next revision, please either (i) have a colleague who is proficient in English and familiar with the subject matter review your manuscript, or (ii) contact a professional editing service to review your manuscript. PeerJ can provide language editing services - you can contact us at [email protected] for pricing (be sure to provide your manuscript number and title). – PeerJ Staff
-
-
-
The manuscript has gotten into good shape. I wish all the best to the authors for their hard work. Thanks
-
-
-
The authors responded to the comments. There are some minor comments:
- The list of abbreviations should be sorted A-Z.
- Figures 8 to 11 should be added in a better resolution. As you are using Python, you can increase the resolution by using the "dpi" parameters from the "matplotlib" Python package.
- The abbreviations in L.528 such as ACC and SEN were not used in the manuscript. You can remove these abbreviations or use them in the tables. If you are going to remove them from this paragraph, then remove them also from the list of abbreviations.
- Add two subsections in the experiments and discussion section. One for the ethics and privacy and another for the relevance of the study from the medical perspective.
-
-
Please read and respond to all three reviews.
**PeerJ Staff Note:** Please ensure that all review and editorial comments are addressed in a response letter and that any edits or clarifications mentioned in the letter are also inserted into the revised manuscript where appropriate.
**Language Note:** The review process has identified that the English language must be improved. PeerJ can provide language editing services - please contact us at [email protected] for pricing (be sure to provide your manuscript number and title). Alternatively, you should make your own arrangements to improve the language quality and provide details in your response letter. – PeerJ Staff
There are some awkward phrasings and repetitive explanations that hinder readability. Language is not clear, intelligible, or professional. Therefore, I recommend that the authors revise the language for conciseness to enhance readability. Figures and tables need clearer captions. The paper is a bit lengthy. It could be more concise. The mathematical notations/statistical analyses are dense. It needs a clearer description and discussion. The introduction is overloaded with background information. Try trimming some unnecessary and repetitive information. I would suggest the author highlight the key contributions of their work more explicitly in the abstract. Try changing the title to this one: "Optimizing CNN Hyperparameters with PSO for Diabetic Retinopathy Detection."
The literature review is limited as the sources are not adequately cited. The authors are advised to cite some more seminal work in this area. PSO and CNNs are well-established techniques. If used, the new adaptive PSO variations or hybrid optimization approaches could have strengthened the work. Nevertheless, this gap can be overcome by discussing some seminal work after 2023. The methodology could have been strengthened further by having more ablation studies. The study lacks practical deployment insights. Applying PSO to existing CNN models limits the novelty of this work. Compare your work with the latest work.
Scalability is limited. How will the issue of scalability to large-scale datasets be verified?
The study lacks citations to more recent works (2023, 2024, 2025) and should compare with other optimization approaches beyond genetic algorithms.
1. Please discuss deployment challenges and future research directions in the conclusion section, as they are not well explored.
2. The novelty of the work is limited. The use of PSO for CNN tuning is not a new technique. Therefore, the authors are advised to provide better novelty justifications and clarity improvements.
3. The conclusion summarizes key findings effectively but lacks strong discussions on future research directions and practical implications. The potential impact of this work is limited outside of DR classification.
This work is highly relevant to AI-driven ophthalmology and medical imaging. The paper also demonstrates improved classification accuracy. The use of a well-established benchmark APTOS 2019 dataset for DR classification is a promising feature of this paper.
The manuscript is generally well-written and written in clear and professional English. The introduction introduces the subject and motivation for optimizing CNN hyperparameters for diabetic retinopathy classification using PSO. However, the background could benefit from including more recent literature (2022–2024) on CNN optimization techniques and hybrid metaheuristic models to better situate the proposed approach within current research trends. The structure conforms to standard norms, but the description of related works and novelty could be made more distinct to highlight the contribution explicitly.
1. Xu, X., Luo, Q., Wang, J., Song, Y., Ye, H., Zhang, X.,... Shi, G. (2024). Large-field objective lens for multi-wavelength microscopy at mesoscale and submicron resolution. Opto-Electronic Advances, 7(6), 230212. doi: 10.29026/oea.2024.230212
2. Cai, L., Fang, H., Xu, N., & Ren, B. (2024). Counterfactual Causal-Effect Intervention for Interpretable Medical Visual Question Answering. IEEE Transactions on Medical Imaging, 43(12), 4430-4441. doi: 10.1109/TMI.2024.3425533
3. Huang, H., Wu, N., Liang, Y., Peng, X., & Shu, J. (2022). SLNL: A novel method for gene selection and phenotype classification. International Journal of Intelligent Systems, 37(9), 6283-6304. doi: https://doi.org/10.1002/int.22844
4. Jia, Y., Chen, G., & Chi, H. (2024). Retinal fundus image super-resolution based on generative adversarial network guided with vascular structure prior. Scientific Reports, 14(1), 22786. doi: 10.1038/s41598-024-74186-x
5. Song, W., Wang, X., Guo, Y., Li, S., Xia, B.,... Hao, A. (2024). CenterFormer: A Novel Cluster Center Enhanced Transformer for Unconstrained Dental Plaque Segmentation. IEEE Transactions on Multimedia, 26, 10965-10978. doi: 10.1109/TMM.2024.3428349
**PeerJ Staff Note:** It is PeerJ policy that additional references suggested during the peer-review process should only be included if the authors are in agreement that they are relevant and useful.
The article falls within the journal's scope and addresses a relevant problem using appropriate methodology. However, several methodological details are missing or insufficiently described:
The specific hyperparameters optimized by PSO for each CNN model (VGG-16, DenseNet-121, MobileNet) should be listed.
To ensure reproducibility, the PSO algorithm needs configuration details (e.g., population size, iteration limits, inertia weight, and velocity update equations).
Although GA is a comparison, further justification for selecting PSO as the optimization method should be discussed with supporting references.
No ablation study or comparison shows the CNN models' performance before and after PSO optimization, which would help quantify the benefits.
The preprocessing steps on the APTOS 2019 dataset and the handling of data imbalance should be elaborated.
Evaluation metrics are appropriate, but justification for their clinical relevance would improve the interpretation.
The findings are promising, especially with the high accuracy and specificity reported for the PSO-MobileNet model. However:
The paper lacks an analysis of computational cost or training time, which is critical for understanding real-world applicability.
The performance comparison with other state-of-the-art methods on the same dataset is missing. Including this would help validate the proposed approach's superiority.
Limitations of the model (e.g., generalizability, overfitting, reliance on PSO parameters) are not discussed.
The conclusion could be strengthened by identifying clear directions for future work and acknowledging the current study's scope limitations.
This is a relevant and technically sound paper with clear objectives and a practical application. Addressing the methodological gaps and citing more up-to-date references would significantly enhance the quality and reproducibility of the work. An open-source implementation or supplementary material (e.g., code/scripts) would further benefit the reader and increase transparency.
- The quality of the figures (e.g., Figures 1 to 10) and their accompanying captions requires improvement, as they currently lack sufficient detail and clarity. Enhance these elements to ensure they effectively communicate the intended information.
- Include a table of abbreviations in the revised manuscript to improve reader comprehension, especially for readers who may not be familiar with all the terms used.
- Add a table of symbols in the revised manuscript to enhance clarity, particularly if mathematical notations are utilized throughout the text.
- Incorporate recent citations from 2022 to 2025 to ensure the manuscript reflects the latest advancements and developments in the field.
- Clearly articulate the research question and identify the gap in the existing literature to ensure the study’s purpose is well-defined and aligned with its objectives.
- Explain what distinguishes the current study from related research to emphasize its uniqueness and contribution to the field.
- Include a comparative table summarizing related studies to highlight the differences and specific contributions of the current work.
- Add a new figure in the introduction section to visually differentiate between the various categories of DR (diabetic retinopathy).
- Confirm in both the text and figure captions that the reported results pertain specifically to the testing subsets to avoid ambiguity.
- Report confidence intervals or standard deviations alongside the results to provide a clearer understanding of the statistical significance and variability of the findings.
- Conduct and include a complexity analysis of the proposed approach to evaluate its computational efficiency and feasibility.
-
-
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.