Next Article in Journal
Valorization of Tomato By-Products: Advanced Extraction Methods and Bioprocessing of Bioactive Compounds and Functional Products
Previous Article in Journal
Maintenance Method for Rail Corrugation Considering Wheel–Rail Interaction Force
 
 
Article
Peer-Review Record

GCS-YOLO: A Lightweight Detection Algorithm for Grape Leaf Diseases Based on Improved YOLOv8

Appl. Sci. 2025, 15(7), 3910; https://doi.org/10.3390/app15073910
by Qiang Hu * and Yunhua Zhang
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Reviewer 4: Anonymous
Appl. Sci. 2025, 15(7), 3910; https://doi.org/10.3390/app15073910
Submission received: 5 March 2025 / Revised: 22 March 2025 / Accepted: 25 March 2025 / Published: 2 April 2025
(This article belongs to the Section Agricultural Science and Technology)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors based upon the YOLOv8 model propose an improved computer vision structure for classification and object detection implemented on images of grape leaf diseases. The proposed structure incorporates to YOLOv8 modules such as Ghost Modules, Reparameterized Convolution (RepConv) modules, Convolutional Attention Modules (CBAM), and a  Shared Convolutional and Separated Batch Normalization Detection Head (SSDetect) module in order to achieve better performance simultaneously reducing the number of parameters and computational cost for real-time application. On the Plant Village Dataset the proposed model achieves superior inference speed and detection accuracy in terms of metrics such as mAP, GFLOPs, and FPS. A comparative analysis and an ablation analysis justifies the improved performance.

The manuscript is well written and the work is well presented.

Please find below some minor comments:

  • In line 120, please explain CBS, C2f, and SPPF abbreviations.
  • In line 135, the C2f-GR module is firstly mentioned. A reference to the figure 7 which follows should be added to facilitate readability.
  • Lines 188 to 191. During training the structure is different than during inference. Please explain how the parameters (i.e. weights) of the inference phase  acquire values since they are not trained.
  • Lines 341 to 344. There is an incorrect repetition.
  • A discussion section is missing even though meaningful discussion is integrated in other sections.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors   Title of the paper:
GCS-YOLO: A Lightweight Detection Algorithm for Grape Leaf Diseases Based on Improved YOLOv8
 
1. What is the main question addressed by the research?
the detection algorithm for grape leaf diseases based on GCS-YOLO, The author got a significant progress in deep learning-based agricultural
disease identification.

2.-  What parts do you consider original or relevant for the field?
The implementation of the three modifications to the YOLOv8:(The Improved GCS-YOLO Object Detection Model)
(1) Replacement of the C2f module with the C2f-GR module,
(2) The integration of CBAM to enable focused attention on critical lesion features and
(3) The detection head is optimized through cross-scale parameter sharing and independent batch normalization layers SSDetect.

3  What specific gap in the field does the paper address?
Detection and classification algorithm for image  

4. What does it add to the subject area compared with other published material?
 In general the main contribution is the improvements for the YOLOv8

5. Please describe how the conclusions are or are not consistent with the evidence and arguments presented.
The conclusion are consistent with the result they show in the modeling and optimization


6. Are the references appropriate?
The references look good and according to the objective of the paper, for the cases as attention mechanism and methods


7.-  General comments:
  a)  it is important to show more information about the data set, for example number of data for the disease 1, the same for the disease 2 and 3,
  b)  are there images without any disease?
  c) the data set is balanced or unbalanced.  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Review the attached file.

Comments for author File: Comments.docx

Comments on the Quality of English Language

The English language must be revised entirely. Correct spelling mistakes, writing, and paragraph length, making sure that they are not too long.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

Manuscript ID-applsci-3540368

The manuscript entitled “GCS-YOLO: A Lightweight Detection Algorithm for Grape Leaf Diseases Based on Improved YOLOv8”


This study investigates the GCS-YOLO, a lightweight grape leaf disease detection algorithm based on an improved YOLOv8. It incorporates the C2f-GR module, RepConv, and CBAM to enhance feature extraction and detection efficiency. The model achieves a 45% reduction in parameters and computational load, with a 1.3% increase in mAP, making it suitable for real-time and edge deployment. Despite the well-executed study, there are areas requiring further clarification to confirm its applicability. However, I have the following comments that will improve the quality of the manuscript.:

 

  1. How does poor image quality, like low resolution, noise, or obstructions, affect the algorithm's accuracy, especially in field conditions?
  2. Does GCS-YOLO lack transparency in its decision-making process, and how could this be a problem if users need detailed explanations for predictions in critical agricultural applications?
  3. The author should define the abbreviation FAO before using it as an abbreviation for the first time, please check for others in the manuscript.
  4. The quality of Figure 10 can be improved as it is difficult to understand.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors heeded my suggestions. Therefore, I accept the article.

Back to TopTop