Computer Vision Framework for Wheat Disease Identification and Classification Using Jetson GPU Infrastructure

Round 1

Reviewer 1 Report

This paper mainly investigated the computer vision framework for wheat disease identification and classification using Jetson GPU infrastructure. However, there are still some major issues that should be resolved and clarified before considering possible publications.

Some specific comments are as follows:

1. There exist some grammar issues including the choices of words, the sentence structure, and the use of articles especially "the". Please further polish the language and correct the grammar errors for a better readability.
2. All the figures are badly designed with low quality. Besides, all the figures are not shown properly.
3. The tables are also not shown properly. It is difficult for reviewers to justify the effectiveness of the paper.
4. If possible, please select more papers published recently for comparison.
5. Please show some visualization results for better understanding the task of this paper.
6. In the RELATED WORK part, a deep literature review should be further given, particularly regarding the backgrounds of image classification and discriminative feature learning. Therefore, the reviewer suggests discussing the advances by citing some references, e.g., “When deep learning meets metric learning: remote sensing image scene classification via learning discriminative CNNs”and “Task-wise attention guided part complementary learning for few-shot image classification”.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments to the Author

This work introduces a deep learning-based approach for disease prediction in wheat crop production and its classification. The author evaluated the model performance w.r.t baseline and compared the state-of-the-art methods. The work has some meaningful results for readers. However, in my opinion, the work needs more improvement in results and evaluation parameter analysis. 

1. Abstract Section can be improved and written better. Example: disease with 99.38 accuracy (no percentage figure)
2. Experimental improvements over the standard baseline methods are satisfied

1. The novelty of the paper is incremental. The work has some meaningful results for future readers. However, in my opinion, the work needs more improvement on model evaluation and result in analysis.

4. The presentation of the paper can be significantly improved if the author can review the paper with an experienced English writer and polish it to make it more readable: such as long sentences, tense, misused punctuations, etc.

Example: Introduction section.

5. The contribution section is Section 1.1 is very basic. A better explanation might help.
6. The reviewer would like to know a more (detailed explanation) about the optimization techniques followed in model training. (Elaborate).
7. The reviewer suggests the author explain the result findings in the quantification section.
8. All the figures are out of the Paper limit and must be an alignment issue. Please rectify those in the revised paper.

References to Author (Classification):

“Adaptive CU Mode Selection in HEVC Intra Prediction: A Deep Learning Approach", Springer, Circuits, Systems, and Signal Processing, vol 38, pp 5081–5102, 2019.

"Cognitive Analysis of Working Memory Load from Eeg, by a Deep Recurrent Neural Network," 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Calgary, AB, 2018, pp. 2576-2580.

Multi-path dilated convolution network for haze and glow removal in nighttime images. Vis. Comput. 2021, 1–14

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

 This work describes the automating the conventional crop disease identification process using a deep learning framework in the case of Ethiopia’s agricultural Sector. Comments to the authors:

1) Figure 1 Wheat Rust Early Warning system requires detailed explanation and it needs to be redrawn.

2) How the data processing has been handled in the paper.

3) There are some typos and grammatical errors.

4) Expand the results section with proper discussion.

5) The results compared with other approaches may be provided.

6) All figures need to be replaced with better quality ones.

Author Response

Please see the attachment

Author Response File: Author Response.pdf