Deep Learning-Based Image Recognition of Agricultural Pests

Round 1

Reviewer 1 Report

-          The abstract should be strengthened.

-          Grammatical errors should be corrected throughout the article.

-          The contribution of this article to future studies should be stated.

-          The conclusion is very inadequate. The conclusion part should be strengthened.

-          After the necessary corrections are made, the article can be published.

Author Response

Hello, dear reviewers! 
I would like to thank you for taking the time out of your busy schedule to review my manuscript. After seeing your suggestions, I have carefully reviewed my paper and revised the English language. The conclusion section has been strengthened, and future research contributions have been described.
Thank you very much for your suggestions.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors designed a new target detection framework based on Cascade RCNN, aiming to solve the problems of large image size, many pest types, and a small and unbalanced number of sample datasets in pest sample datasets. Specifically, this paper performs data enhancement on the original samples to solve the problem of a small and unbalanced number of examples in the dataset and develops a sliding window cropping method, which can increase the perceptual field to learn sample features more accurately and in more detail without changing the original image size. Secondly, combining the attention mechanism with the FPN layer enables the model to learn sample features that are more important for the current task from both channel and space aspects.

I am convinced with the work done by the authors, their presented some innovation. But some minor corrections or suggestions are as under.

1. You should considerably improve the language of your paper.

2. Avoid the use of "we".

3. At the end of the Introduction, please delete the sentences from "the main contributions of this paper are as follows........" Only mention objectives of the study not contributions.

 

Author Response

Hello, dear reviewers! 
Thank you for taking the time out of your busy schedule to review my manuscript. After seeing your suggestions, I have carefully reviewed my paper and revised the English language to avoid using "we" as much as possible. The end of the Introduction has also been changed.
Thank you very much for your suggestions.

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript “Deep Learning-Based Image Recognition of Agricultural Pests” deals with target detection framework based on the Cascade RCNN. The manuscript needs to be improved a lot before the final decision is made. The authors tried to put all data, but the analysis and results were not discussed in detail. This need to be revisited. Moreover, the manuscript lack several references, which need to be cited accordingly. Kindly also look into the language and framing of sentences which need further clarification.

 

Comments

·       It is not clear whether the manuscript is research or review. Kindly mention this in the MDPI template.

·       LN 15-17: Please rewrite the line.

·       LN 18 and 24: Please define the abbreviation

·       In the abstract section, please add some data which is derived from the experiment. It seems to be only the conclusion of the study.

·       LN 43: Is this amount “$70 billion” of the world or of a particular country? Please specify.

·       LN 47: “Headache”, Please don’t use such term. Replace with better phrase or word.

·       LN 47-53: Please cite some recent study here

·       LN 57: What is CNN? Please define the abbreviation. And also for R-CNN

·       LN 115-117: Please rewrite the line.

·       Why “two-layer neural network (MLP) with the number of neurons in the first layer as C/r”? Please explain.

·       Which crop is used in the experiment? Kindly specify whether the single crop is used or multiple crops are used.

·       LN 199-202: Please rewrite the line.

·       Fig. 5: There is no scale used in the photograph. Kindly provide the scale for the image in the rest of the photograph also.

·        Fig. 7 Need to be explained in detail.

·       The caption of figure 8 needs to be explained in detail. It's not clear.

 

·       The conclusion section needs to be revisited. Also, add future thrust in this section. 

Author Response

Hello, dear reviewer! Thank you for taking the time out of your busy schedule to review my manuscript. After seeing the suggestions you gave me, I have carefully reviewed my paper and made the following changes.

1. The manuscript is a research paper。
2. I define abbreviations such as CNN as well as RCNN.
3. In the summary section, I added data from the experiment to make it look more adequate.
4. I have carefully revised the sentences that should be rewritten.
5. The first neural unit of the double-layer neural network (MLP) is C/r, the best case of the attention mechanism studied by researchers.
6. Multi-cropping is used in our model.
7. Fig. 5 is only a picture in the data set that is changed to 1280*1280, obtained by data enhancement, and then reduced and put into the manuscript.
8. Fig. 7 I reinterpreted it to make it look clearer.
9. The title of Fig. 8 means displaying some sample pictures in the data set.
10. Part of the conclusion has been revised, and future work has been added.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

The authors made significant changes in the manuscript. I congratulate the author for this work.