Enhancing Rice Crop Management: Disease Classification Using Convolutional Neural Networks and Mobile Application Integration

Round 1

Reviewer 1 Report

The author introduces a new technology of early rice disease detection based on K-means clustering algorithm and Convolutional neural network (CNN) combined with image processing. This technology has a certain degree of innovation, and the author designed some experiments to verify its ability to detect and classify brown spot, bacterial leaf blight, and leaf smut.

Modification suggestions:

 

1. It is recommended to use ablation experiments to verify the effectiveness of k-means clustering algorithm, image masking, and threshold processing;

2. It is recommended to use newer literature for comparison with existing work and increase the proportion of cited latest literature;

3. It is recommended to divide the dataset in more detail, such as the number of images of bacterial leaf blight, brown spot, and leaf smut.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

I agree that it is very important to detect rice disease in the early stage and start disease treatment ASAP.  Image processing is an adequate method. The problem with your paper is that it is not a new finding, and many papers cover this subject. It is hard to find out the novelty of your work. Therefore I suggest rewriting the paper to indicate your novel contributions clearly. In addition, please stick to the rule when using the acronym for the first time; it needs to be explained. An example is on line 106 of CCNN. Section Author Contribution:  is messy. Nevertheless, I enjoyed reading your paper.

The quality of the English language is fine

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The authors have used image processing and machine learning techniques to detect and classify rice diseases, which is a promising approach with significant implications for agricultural productivity and food security. This work presents a novel technique that combines the K-means clustering algorithm and Convolutional Neural Network (CNN) for the detection of rice diseases. However, the manuscript in its current form is not suitable for publication in this journal.

1.      One major deficiency of this work could be the limited reproducibility of the proposed technique. The paper states that the model has been trained and tested using a specific dataset comprising 900 images, with 60% for training and 14% for testing. However, evaluating the model's performance on a more diverse and representative dataset is crucial to assess its robustness and applicability in real-world scenarios.

2.      It is crucial to contemplate that rice diseases can exhibit variations in symptoms and appearances depending on various factors, including geographic region, climate, and rice cultivar. Therefore, it is essential to validate the proposed technique on datasets that cover a wide range of these factors to ensure its effectiveness across different regions and rice varieties.

3.      The structure and writing style of the paper deviates from the conventional scientific manuscript format, resembling more of a thesis section rather than a journal manuscript. To enhance the overall quality of the content, I recommend observing previously published papers in this field and incorporating their established conventions into the writing.

4.      Proper citations of the facts and previous studies are very important and are completely neglected to start from the introduction to the discussion.

 

 

 

Minor corrections are required

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

All the comments were properly addressed.