Next Article in Journal
A Field-Programmable Gate Array-Based Quasi-Cyclic Low-Density Parity-Check Decoder with High Throughput and Excellent Decoding Performance for 5G New-Radio Standards
Previous Article in Journal
RETRACTED: Galeoto et al. Assessment Capacity of the Armeo® Power: Cross-Sectional Study. Technologies 2023, 11, 125
 
 
Article
Peer-Review Record

Rice Leaf Disease Classification—A Comparative Approach Using Convolutional Neural Network (CNN), Cascading Autoencoder with Attention Residual U-Net (CAAR-U-Net), and MobileNet-V2 Architectures

Technologies 2024, 12(11), 214; https://doi.org/10.3390/technologies12110214
by Monoronjon Dutta 1, Md Rashedul Islam Sujan 1, Mayen Uddin Mojumdar 1, Narayan Ranjan Chakraborty 1, Ahmed Al Marouf 2,*, Jon G. Rokne 2 and Reda Alhajj 2,3,4
Reviewer 1:
Reviewer 2: Anonymous
Technologies 2024, 12(11), 214; https://doi.org/10.3390/technologies12110214
Submission received: 13 September 2024 / Revised: 23 October 2024 / Accepted: 25 October 2024 / Published: 29 October 2024
(This article belongs to the Section Information and Communication Technologies)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper at hand proposes a method for rice leaf disease classification from RGB images employing image processing techniques, i.e., bilateral filtering, and deep learning models. The paper is well-written and organized. The method demonstrates some novelty in the field leading to promising recognition results. Both k-fold cross-validation and comparative analysis are presented. Hence, the experimental study is sufficient to highlight the main contributions of the presented work. Yet, to the reviewer’s point of view, some minor changes should be made.

1. In both abstract and the main body of the manuscript, the use of the term CNN for representing the proposed model is misleading. For instance, in the abstract it is stated that “the study employs advanced neural network architectures for classification, including Cascading Autoencoder with Attention Residual U-Net (CAAR-U-Net), MobileNetV2, and Convolutional Neural Networks (CNN).” Yet, both MobileNetV2 and U-Net constitute CNN architectures, as well. Hence, referring to the proposed model simply as CNN can cause misunderstanding to the reader. Please, suggest a distinctive term for referring to the proposed model, such as, Ours, Proposed CNN, Custom CNN, etc.

2. All abbreviations, like Convolutional Neural Networks (CNN), should be explained again at their first appearance in the main body of the manuscript. Then, only the abbreviation (e.g. CNN) needs to be referred. Please revise carefully.

3. In lines 129-130: “Deep learning, a branch of machine learning, can be used to diagnose plant diseases from complex visual input” please cite works in deep learning for image classification below:

LeCun, Yann, Yoshua Bengio, and Geoffrey Hinton. "Deep learning." nature 521.7553 (2015): 436-444.

Kansizoglou, Ioannis, Loukas Bampis, and Antonios Gasteratos. "Deep feature space: A geometrical perspective." IEEE Transactions on Pattern Analysis and Machine Intelligence 44.10 (2021): 6823-6838.

4. Both MobileNetV2 and CAAR-U-Net architectures are well-established pre-trained feature extractors for classification tasks. The authors could discuss further how they explain the improved performance achieved by the shallower and more customized CNN model proposed. Is there any correlation between the model and the dataset provided? Please, discuss further.

5. In the conclusion, the authors are encouraged to present their ideas for future work or practical applications.

Author Response

Thank you so much for taking time to review the paper and suggesting the changes. We have considered all your comments and updated the paper accordingly. The changes are highlighted on the paper.

Please find the detailed responses in the attached file. 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dutta et al. deal with recognition of leaf diseases based on internet source-obtained images and Convolutional Neutral Network (CNN) models. Authors are encouraged to deal with the following comments.

 

Line 23: “caused” in place of “affected”

Lines 27, 29, 32, 47: Latin names in italics

Line 33: please provide the virus name

Lines 37, 44, 52 and throughout the manuscript: the first name of the author is not cited within the text

Line 47: why provide here the Latin name, and not the first time that the species is mentioned (Line 19)

Line 132: why provide here the abbreviation, and not the first time that the term is mentioned (Line 50)

 

Introduction is too long, and contains several details. You need to start more general and then focus on disease detection.

In the last decade, CNNs have been increasingly employed in plant phenotyping community. They have been very effective in modeling complicated concepts, owing to their ability of distinguishing patterns and extracting regularities from data. Examples range from variety identification in seeds (Taheri-Garavand et al., 2021 Plants 10, 1406) up to leaf water status information (Fanourakis et al. 2024 Plant Growth Regul 102, 485496).

Then, you better present a Table (like the Table 6 of this study) providing the key findings focusing on leaf disease detection. Since so many papers focus on rice, is it NOT really needed to mention other species.

 

Materials and Methods

Impressing that all the images were obtained from internet. Is there any related information regarding cultivars, growth conditions, disease severity, light conditions during image acquisition? Please provide more representative pictures per disease in Figure 2.

Please indicate the spectrum, where images were obtained? visible portion of the electromagnetic spectrum (400–700 nm)?

How many cultivars? How many images per cultivar under study?

 

Discussion is actually missing! I suppose it is section 4.4 (line 343), which is rather short. You need to compared your findings with previously-obtained ones, and highlight what is new and novel

Can the present algorithm applied in fruit harvested in other seasons? Or the same process ought to be performed again? What is the value of the obtained model then?

Enrich the discussion by including: ”On a commercial scale, evidently, a capital investment is initially required for adopting the employed approach (Taheri-Garavand et al., 2021 Industrial Crop Prod 171, 113985). Nevertheless, the wide-ranging large-scale commercial applications can provide high returns through considerable improvements in shorting process enhancement and cost reduction.”

Comments on the Quality of English Language

minor english changes are needed 

Author Response

Thank you so much for taking time to review the paper and suggesting the changes. We have considered all your comments and updated the paper accordingly. The changes are highlighted on the paper.

Please find the detailed responses in the attached file. 

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

my comments were adequately addressed

Comments on the Quality of English Language

minor changes needed

Back to TopTop