Fast Self-Attention Deep Detection Network Based on Weakly Differentiated Plant Nematodess

Round 1

Reviewer 1 Report

Page 9: where N is the number of blocks, typo, make N italic

The results of this method are very impressive: around 100% (classification accuracy) and zero annotation time etc., Can you please reiterate one more time precisely why you got such high accuracy as compared to other models (YOLO, Faster RCNN, etc.? How can you assure other researchers there is no overfitting in this result?

Let's talk about the zero-annotation time: under what condition authors reiterate zero annotation time? There are 2111 samples in testing. Can you please include the testing time in revise version of paper?

Discussion:

Semi-supervised learning: finding class label of large unlabeled samples based on minimum numbers of label samples. Does this method work resemble the semi-supervised learning algorithm?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This  Paper is better defined and the Parameter Setting is good.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Following are my comments to further improve this interesting work on fast self-attention deep detection network:

·        The abstract is written point wise like a, b, c. It is bit confusing. Kindly write as a single paragraph of text.

·        Include at least three more keywords to improve visibility.

·        Improve the link between state of the art and objectives of the paper.

·        A small table listing contemporary work will improve the visibility of the paper. You may include and discuss similar papers such as 10.1080/15567036.2022.2067268, 10.1016/j.asoc.2020.106495, 10.1109/TGRS.2021.3073562, 10.1016/j.seta.2022.102708

·        Did you carried out the uncertainty analysis.

·        Figure 3 is not clear.

·        Conclusion should be improved using bullet wise results.

·        The conclusion should have both quantitative and qualitative results.

It is well written and compact study. It may be considered after modification.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I have no more comment!!!!

Author Response

Thank you for your suggestions on revising the article.

Reviewer 3 Report

The author should work on point 3 and point 4. 

My query:

Point 3 of initial report is not answered properly.

Point 4 of initial report is included in the manuscript. This suggestion should be added to the list of references.  

Author Response

Point 3:  Improve the link between state of the art and objectives of the paper.

Point 4:  A small table listing contemporary work will improve the visibility of the paper. You may include and discuss similar papers such as 10.1080/15567036.2022.2067268, 10.1016/j.asoc.2020.106495, 10.1109/TGRS.2021.3073562, 10.1016/j.seta.2022.102708

Response 3 and 4：In order to improve the link between state of the art and objectives of the paper, I listed the existing research methods on the second page. The specific references are 23-26. The specific modification details will be shown in the revised version. Thank you for your valuable suggestions for the article.