A Ginseng Appearance Quality Grading Method Based on an Improved ConvNeXt Model

Round 1

Reviewer 1 Report

The paper presents a correct methodological approach and shows interesting results. Although the application is limited to a very specific case, the method could be tried on different case studies.

However, I have some suggestions and requests for clarification which could be useful to further improve the paper.

11)      The methods used for expanding the dataset are described three times in the paper: (lines 132-144, 267-269, 303-310). I think that one time is enough.

22)      In equation (1) define I and F.

33)      Section 3.2.3 should be harmonized with what is stated at the beginning of Section 4.4. In fact, when you write in 3.2.3 that "this study utilizes the PReLU activation function," it seems that you used only that, whereas in 4.4 it turns out that you also used GELU, ReLU and even sigmoid. Furthermore, in 3.2.3 you should mention the difference between ReLU and GeLU, possibly integrating the graph in figure 5.

44)      Specify if Table 5 refers to training or test.

55)      Lines 311-314. Check the sentence. I understand that 7.3% is 85.97-78.67, but the sentence is not clear.

66)      Line 320. Which four types? We have 8 models. If you want to compare NUM8 with NUM3, NUM5 and NUM6 you have to write “with the other 3 types”

77)      Lines 332 and following. You state that re-parametrization module is more effective compared with other modules. However, the combination of the PReLU and re-parametrization has the lowest effect on the accuracy rate. These two sentences seem in contradiction.

88)      I suggest to delete figs 7 and 8, and the respective comment. In my opinion they are not relevant.

99)      A single little paragraph  for Discussion is a nonsense. I suggest to integrate it in results.

Some sentences are not clear.

Author Response

For more information, please refer to the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The author(s) did a great works introducing a Ginseng quality grading method based on ConvNext model. However, the following points are required to revised:

• References used are related, well utilized, and update. 

• Define abbreviation before its use (i.e.GELU, PReLU, …etc)

• Add aparagraph at the end of Introduction section to characterize the manuscript structure.

• Section 4 Experimental Validation and Analysis of the Results plus Section 4 Discussion, kindly elaborate your discussion- merge analysis with discussion- and renumbering your manuscript.

• Revised the Conclusion section to only summarizes the work and outlines the futuristics.

I can accept the work after amendment,

Author Response

For more information, please refer to the attachment.

Author Response File: Author Response.docx