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.
The manuscript may be accepted.
[# PeerJ Staff Note - this decision was reviewed and approved by Xiangjie Kong, a PeerJ Section Editor covering this Section #]
-
-
-
Thank you for your submission. Please revise and resubmit according to the reviewer comments.
**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.
**PeerJ Staff Note:** Your submission appears to have been at least partially authored or edited by a generative AI/large language model. When you submit your revision, please detail whether (and if so, how) AI was used in the construction of your manuscript in your response letter, AND mention it in the Acknowledgements section of your manuscript.
Thank you for the opportunity to review the manuscript titled "Advancing Crop Health with YOLOv11 Classification of Plant Diseases." The study presents a well-structured analysis of using YOLOv11 for plant disease classification. The authors demonstrate a strong understanding of the subject and provide sufficient background information. The manuscript is written in professional and clear English, making it accessible to an international audience. The abstract follows the expected structure, covering context, objectives, methods, results, and conclusions. Figures and tables are relevant, well-labeled, and effectively illustrate the findings. The introduction includes a comprehensive literature review that establishes the study’s relevance.
While the manuscript is well-prepared, a few areas could be improved. The authors should ensure that all figures and tables are uploaded separately, following PeerJ submission guidelines. Additionally, the introduction and related work sections would benefit from more references to recent literature, particularly studies published after 2020. The discussion could also be expanded to emphasize the practical implications of the findings beyond academic research, especially in relation to agricultural policymakers and industry practitioners. Furthermore, the limitations of the study should be more explicitly highlighted, either in the discussion or the conclusions section.
The research aligns well with the aims and scope of PeerJ Computer Science and presents a strong experimental design. The research questions are clearly defined, addressing a relevant knowledge gap. The dataset is sufficiently large, with 2,904 images properly divided into training, validation, and test sets. The methodology is well-documented, making it possible for other researchers to reproduce the study. Ethical standards are maintained, and a data availability statement is included.
To further strengthen the methodology, it would be beneficial for the authors to provide a more detailed discussion of the data preprocessing steps, particularly regarding how they handled class imbalances, if relevant.
To further strengthen the methodology, I suggest about Data Preprocessing: Provide more details about how class imbalances were handled (if applicable) and any data augmentation techniques used. For Computational Requirements: Include details about the hardware setup (e.g., GPU, memory, processing time) used to train and test the YOLOv11 model. This would help readers assess the feasibility of implementing the model in practical scenarios and for Model Justification: Expand on why YOLOv11 was chosen over other deep learning models. A clearer justification of its advantages over alternatives (e.g., YOLOv8, EfficientNet, ResNet) would strengthen the manuscript.
The results are presented clearly and align well with the study's objectives. The conclusions are well-supported by the findings, and the statistical analysis is thorough. The authors effectively demonstrate the model’s effectiveness using accuracy, precision, recall, and F1-score, reinforcing the study’s credibility. The discussion effectively situates the findings within the broader context of prior research.
To further enrich this section, consider the Dataset Variability: Address potential challenges related to dataset variability, such as differences in environmental conditions, lighting, or image sources, and their impact on model performance and for Future Research Directions: Discuss potential improvements to the model or real-world implementation challenges that need to be addressed in future work. Moreover, consider the Bias and Generalizability: Acknowledge any possible biases in the dataset and their implications for the practical deployment of the model.
Overall, this is a well-structured and methodologically sound manuscript with strong potential to contribute to the field of plant disease classification using deep learning. With some refinements, the impact of the study can be further enhanced.
The manuscript is suitable for publication following minor revisions, including ensuring all figures and tables comply with PeerJ's submission format, expanding the discussion on practical agricultural applications, providing more details on data preprocessing and computational requirements, addressing potential dataset biases, and clarifying the choice of YOLOv11 over other deep learning models. With these improvements, the study will offer a more comprehensive and impactful contribution to plant disease classification research.
The manuscript is generally well-written in clear, professional English, though there are some minor grammatical errors and typographical inconsistencies that should be addressed. The introduction provides adequate context for the work, explaining the importance of plant disease detection in agriculture and the potential benefits of automated systems.
The literature review is comprehensive, covering relevant prior work in deep learning for plant disease classification. Table 1 provides a good summary of related studies, their methodologies, and results. The authors have done a commendable job of organizing the paper into a logical structure following standard scientific reporting conventions.
The figures are relevant and informative, particularly Figure 4, which showcases sample images of different plant diseases. However, some figures (especially Figures 2 and 3) would benefit from higher resolution and clearer labelling. Additionally, Figure 9, showing the learning curves, could be improved with more descriptive captions explaining the specific metrics being displayed. In addition, the confusion matrix within the figure is not legible (consider splitting Figure 9 into sub-figures and labelling the axes for the figures).
Other comments:
1.1 - 1.4 can be consolidated - there is no need for subheadings
The research addresses a relevant question in agricultural technology: improving plant disease classification using deep learning. The authors clearly identify the knowledge gap they aim to fill by highlighting limitations in existing approaches, such as computational intensity and scalability issues.
The methodology is described in sufficient detail, with a clear explanation of the dataset (2,904 images across 12 disease classes), model architecture (YOLOv11), and evaluation metrics. The structure of the YOLOv11 model is explained thoroughly, including the backbone, neck, and head components, along with key architectural innovations like the C3k2 block and C2PSA spatial attention mechanism.
The hyperparameters used for training are well-documented in Table 3, and the evaluation process is clearly outlined in the pseudocode in Figure 6. The dataset split (70% training, 20% validation, 10% testing) follows standard practices in machine learning.
Areas that should be further improved:
- More details on computational resources used for training would be helpful
- The choice of YOLOv11 over other architectures could be more explicitly justified
- A more in-depth explanation of why specific hyperparameters were chosen
The results presented in Tables 4-7 are comprehensive, showing the confusion matrix, classification report, overall performance metrics, and per-class metrics. With an accuracy of 96.90%, precision of 97.08%, recall of 96.90%, and F1-score of 96.89%, the model demonstrates strong performance across the 12 disease classes.
The authors appropriately discuss their results in the context of other recent methods (Table 8), showing comparable or superior performance. The discussion section thoughtfully addresses limitations and potential future directions, including the need for dataset expansion, environmental robustness, and real-world deployment considerations.
The conclusions are directly linked to the original research question and are supported by the presented results. The authors don't overstate their findings and acknowledge areas for future improvement.
Further comments
- A more detailed analysis of misclassifications would strengthen the paper
- Additional ablation studies to quantify the contribution of specific architectural components (this is strongly recommended since only YOLOV11 is presented without any other comparison)
- The authors should discuss potential biases in the dataset and how they might affect real-world applications
1) This paper talks about the advancement of crop health with YOLOv 11 classification of plant diseases.
2) This study explores the efficacy of a YOLOv11-based deep learning model for classifying plant diseases in a multi-class setting, focusing on diseases affecting beans, strawberries, and tomatoes. The model was trained and evaluated on a dataset comprising 2,904 images across 12 distinct disease classes. The results demonstrate that YOLOv11 delivers strong performance, achieving an accuracy of 96.90%, precision of 97.08%, recall of 96.90%, and an F1-score of 96.89%. Additionally, the model’s total inference time of 7.99 seconds highlights its computational efficiency, making it well-suited for real-time or near-real-time applications in precision agriculture. The YOLOv11-based framework developed in this study provides a scalable, accurate, and efficient solution for plant disease classification. By leveraging advanced deep learning techniques, the model overcomes key limitations of traditional and existing automated systems, such as restricted disease coverage, computational inefficiency, and scalability challenges.
3) This paper has potential. However, it requires a major revision.
4) The keywords should be improved to attract more readers.
5) The first 2 lines of your abstract should be defined with a clear problem statement of why this work is proposed. This is currently missing in the abstract.
6) The introduction section of this work needs fine-tuning. The details of elements missing in the introduction section can be seen in the following comments.
7) What is the contribution of the paper? It must be clearly mentioned in the introduction. So far, these elements are not obvious in the introduction.
8) What is the motivation of the paper? It must be clearly mentioned in the introduction. So far, these elements are not obvious in the introduction.
9) What is the scope of the paper? It must be clearly mentioned in the introduction. So far, these elements are not obvious in the introduction.
10) What are the preceding affined works in this research area? A table shall be drawn showing the limitations of the previous works in this area, and what new information are you bringing for the readers? Or is it yet another technical paper? Please ensure that Table 2 you have presented covers all the elements as suggested.
11) How can a reader reproduce your work? This is the most important question to be answered. Since there is limited formulation in your work, there should be a pseudo-code to elaborate on the proposed scheme before Section 4 on results and analysis. In this work, a possible pseudocode could be designed for practitioners to understand things smoothly. The variables of input and output are required to be defined in the pseudocode. And then that defined language is required to be used in the rest of the lines. Please follow the above steps accordingly.
12) In the results and discussion section, what mainstream techniques are utilized for comparison? A graph/plot/tabular representation should be made. The superiority of your proposed scheme shall be mentioned in the conclusion based on the outcome of the comparison with other mainstream techniques.
13) For a paper talking about advanced crop health, elements like smart agriculture, which can safeguard the food and enhance the process, should be mentioned in the introduction. This will allow for widening the spectrum of audience towards irrigation systems, food security, and smart agriculture.
14) Format of references requires fine-tuning. In some of the references, the title is not mentioned at all. In some references, the page number is reflected at the end, and in others, the year is reflected at the end. Similarly, in some references, the title is in inverted commas, and in some it is now. Also, in some, the first letter of the title of the paper is only capitalized. In others, all first letters of the titles of the paper are capitalized. Please maintain consistency.
-
-
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.