Locality Preserved Selective Projection Learning for Rice Variety Identification Based on Leaf Hyperspectral Characteristics

Round 1

Reviewer 1 Report

Comments

Introduction

The introduction should only set out the research in relation to the current study, state the problem or shortcomings of previous research and finally state the objectives.

In the case of this article, the authors have failed to structure the introduction properly. For example, from line 99 to line 109, this paragraph would be better placed in the method section. Then, from line 111 to line 116, it should really be in the results section.

Materials

It is preferable to present the soil characteristics in table form (Lines 121 to 127).

In the experimental design section, it is better to use the word transplant rather than insertion.

Line 142: At the heading stage

How did you normalize the data? (Line 193)

General questions:

1. What is the concept behind "Locality Preserved Selective Projection Learning," and how does it contribute to the accurate identification of rice varieties based on leaf hyperspectral characteristics?

2. Could you explain the significance of hyperspectral characteristics in distinguishing between different rice varieties? How does hyperspectral imaging capture unique features of rice leaves for identification purposes?

3. In comparison to traditional methods of rice variety identification, how does the "Selective Projection Learning" approach enhance accuracy and efficiency? Are there specific scenarios or conditions where this approach excels?

The entire article must be edited by a native English speaker.

Author Response

Thanks for your helpful comments. Please check the attachment to see the detailed response.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors presented an interesting and promising work on rice species identification with hyperspectral information and LPSPL as feature extraction method. However, I have some concerns to be clarified after the paper can be accepted for publication:

• An image for each of the varieties and growth stages (20 images) would better describe the trickiness of the problem. Please consider adding them.

• Is the dataset publicly available?

• Why don’t you evaluate techniques like t-SNE or Umap? Theoretically, they are more promising than PCA. 

• Is there any GitHub repo with the source code?

• Why didn’t you use more ML algorithms besides SVM? This way, you can claim that LPSPL boosts the performance of different families of classifiers.

• What are the next steps after this research?

Author Response

Thanks for your constructive comments. The detailed response is submitted in the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear Authors,

Thank you for providing materials about rice variety identification. I found this paper very interesting. By the way, the research previously was done by same authors providing different methods of identification and is not included in reference list. Therefore, I have a few questions.

1. The paper named "Rice Variety Identification Based on the Leaf Hyperspectral Feature via LPP-SVM" consist of the same dataset as in this paper, is it right? I found part of Figure 1 is the same one from previous paper.

2. Why you did not mention your previous research paper which was done on LPP method?

Abstract - I suggest revising abstract with more details about explanation of issue in this niche. What is the aim to find rice variety based on leaf hyperspectral imagery?

Introduction - Recheck references.

49-59- Are these all related to potato research?

80 - References.

95 - References.

Figure 3 - Figure 3 is controversial to understand. Reconsidering representation of it.

Table 3 - Please, revise this table.

310-324 - It is only one paragraph explaining 4 figures and 4 tables. Please consider adding more details.

Discussion and Conclusion - There is no discussions. Consider separating those chapters.

I suggest revise Introduction and conclusion section. Moreover, it lacks literature review on this topic. Please consider explaining previous research since the data preparation is very similar and looks like a self-plagiarism. Text of figure's axis is very small. It is difficult to read figures and understand them.

Thank you.

Author Response

Thank you very much for your comments. The detailed response is presented in the attachment. Please check the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

No more comments 

Author Response

Thank you very much for your constructive comment that improved this manuscript.

Reviewer 2 Report

Thanks for addressing the comments. Some minor issues should be taken into consideration before publication:

- L74: “and so on” -> “among others”.

- Figure 4, 5, 6, 7: It would be preferable to improve the resolution.

- L310: space after commas.

- L457: “shows” -> “show”

Reviewer 3 Report

Dear Authors,

Thank you for the revised materials and detailed explanations.

337 - Please, the "table 4" should appear before the table itself in the text. Also, its Table 4, not TableS 4.

Please, revised the text one more time. I understand that there might be a deadline for some reason. But be sure there is no minor mistakes like above before publishing.

Thank you.

Author Response

Thanks for your suggestions that significantly improve the quality of this manuscript.

As suggested, in line 337, the " Table 4" has been moved to the front of the table itself and the " TableS 4" has been corrected to " Table 4".

Moreover, we have carefully checked and corrected the mistakes from grammar, spelling, and so on throughout the manuscript.