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.
Although one reviewer rejected the manuscript, it contains valuable information, and the authors have addressed all of the reviewers' comments to the best of my knowledge.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
A reviewer suggests rejection, but the other reviewer considers your work slightly more positively. I suggest that you please revise the manuscript thoroughly, paying attention to both reviewers. English grammar and misspellings should be avoided.
**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 Academic Editor 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
This study introduces a novel deep-learning framework that combines advanced image preprocessing, segmentation, and optimization techniques to accurately diagnose and classify rice leaf diseases. The framework utilizes a new clustering method called Gravity Weighted Kernelised Density Clustering (GWKDC) to segment diseased regions effectively, as well as a fine-tuned EfficientNetB0 model for deep feature extraction. To improve convergence and training efficiency, the study proposes an optimization algorithm—Tent Chaotic Particle Snow Ablation Optimizer (TCPSAO). The framework was evaluated on two public datasets (Mendeley and Kaggle), achieving classification accuracies of 98.87% and 97.54%, respectively. The proposed method outperformed other deep learning models and optimization techniques, proving its effectiveness and robustness in real-world agricultural applications.
The experimental design process is unclear because it is not explained why two separate data sets are used. In addition, the evaluation and organization of the experiment and the results do not compare the optimization of the method.
This finding is unclear when many segmentation and optimization methods have been compared on rice datasets. Still, it has not been thoroughly addressed in related studies or comparisons of methods.
However, there are several shortcomings in the current version that the authors should address:
1) The proposed approach combines optimization algorithms and segmentation methods. However, since both techniques are integrated in the experimental phase, it is unclear whether the performance improvements are due to the optimization algorithm or the segmentation process.
2) The comparison in Tables 2, 3, 4, 6, 7, and 8 is not entirely clear, as the experimental scenarios implemented by the authors are not well described, and the data are presented separately, making comparison difficult.
3) Although the model achieves high accuracy, important computational metrics such as training time, model size, and inference latency are not clearly reported.
4) The model is trained on two specific datasets. There is no evaluation of its robustness when applied to noisy images, low-light conditions, or previously unseen diseases.
5) The model is entirely trained and evaluated on pre-collected datasets (Kaggle and Mendeley). However, the authors have not explained why the two datasets are used separately or why they are not combined into a single dataset for training and testing.
6) In the Conclusion section, the authors mention plans to deploy the model on mobile devices in the future. However, they have not assessed the feasibility of such deployment based on the current study.
7) Both segmentation techniques and optimization algorithms have been widely studied and compared in previous research. Therefore, the authors should explore existing limitations in these areas and focus more clearly on addressing them in their study.
• This paper has an interesting contribution in rice leaf disease classification using clustering and transfer learning methods
• The introduction and literature review are quite clear in describing the background of the problem, the importance of the research, and the proposed method.
• The use of English in the paper is quite good academically
1. Add performance analysis of the segmentation method
2. Clarify the amount and distribution of training and test data
3. Provide a segmentation algorithm
4. Show visualization of the segmentation results to demonstrate the effectiveness of the clustering method
1. The results are impressive, especially the high AUC values, which indicate that the model is highly capable of distinguishing between diseased and healthy rice leaves.
2. It would be beneficial to include more detailed information on the model's performance under various conditions (e.g., varying image quality or environmental factors).
Correct any spelling errors, especially those related to equations and their explanations.
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.