DCF-Yolov8: An Improved Algorithm for Aggregating Low-Level Features to Detect Agricultural Pests and Diseases

Round 1

Reviewer 1 Report

The agronomy-2531727 with the title of DCF-Yolov8: An improved algorithm for aggregating low-level features to detect agricultural pests and diseases needs significantly improvements before it can go for further step:

The ms does not have line number to make the life easier for the reviewers!

The Keywords should be revised and the authors should avoid using the same words from the title.

In the introduction section, the authors should add additional references and make the introduction as story and avoid making jumping among different paragraphs.  

L1-2 in introduction: Authors should cite this study: Bio-ecology and the management of Chenopodium murale L.: A problematic weed in Asia

The authors should add suitable citations for this text “With the advancement of deep learning technology, it has become increasingly feasible to learn complex feature representations and leverage the strengths of deep learning in various domains such as image, video, and sensor data analysis. Suitable deep learning algorithms can enable precise identification and classification of crop pests and diseases, assisting farmers in determining the type of pests/diseases and implementing appropriate prevention and control measures.”

L2-3 The authors should cite this text using “Induction of Systemic Resistance in Hibiscus sabdariffa Linn. to Control Root Rot and Wilt Diseases Using Biotic and Abiotic Inducers”

L3-4 Here also authors should add citation for this text and can use: Activity of Essential Oils and Plant Extracts as Biofungicides for Suppression of Soil-Borne Fungi Associated with Root Rot and Wilt of Marigold (Calendula officinalis L.)

Citations within the is is wrong, the authors should go through their ms and correct all these issue such as A. Chaudhary et al. [1] or A.K. Singh et al. [2] or others. Authors should remove the abbreviations for the first or middle names. Check this in whole ms.

The aim of the study at the end of the introduction should be revised because it is not well written as well as the authors should reduce the focus on the related work in section 2.

Some Figures should be improved in terms of numbering different pictures within same Figure.

Some Figures such as Figure 13 is low quality, check this in whole ms please.

Title of the Figures should be improved; they are not well written in details.

Figure 17: Why authors do not make pictures bigger size ?  Please make it bigger, so it will be clear for readers.

The discussion section is very poor written and the authors did not cite any single investigation, this section should be written again.

References: Authors should follow the format of the journal, for example, some format should be bold and others should be italics.

 

 

 

 

Minor editing of English language required

Author Response

Thank you for allowing a resubmission for our manuscript, with an opportunity to address the reviewers’ comments.

We would like to express our gratitude for granting us the opportunity to resubmit our manuscript and address the valuable comments provided by the reviewers. We have attached the following documents.

Author Response File: Author Response.docx

Reviewer 2 Report

In this manuscript, the authors proposed a pest and disease detection method based on YOLOv8. Several modules like CBM and D2F were introduced to boost performance.

 

Here are some comments that the authors should address during revision.

 

(i)               The authors should re-organize the abstract. The research gap and the motivation were unclear. To be exact, what was the challenge in the pest and disease detection task and how did the authors solve this challenge?

(ii)              In the Introduction section, the authors review some recent work about pest and disease detection models, but failed to identify the research gap. As the authors mentioned some of them had achieved great performance, like reference [4], the proposed model achieved a precision of 98%. So, how could this model not be applied to the IP102 dataset? The authors should clearly mention the shortcomings of current models and thereby highlight the novelty and contributions of their own.

(iii) At the end of the Introduction section, the authors should summarize their main contributions to the field in bullet points.

(iv)            The organization of Section 3 is not good. As the title suggested, materials and methods. Where are the materials, especially for the data. Also, it is hard to follow the authors’ overall idea by beginning with the activation function. It is important to present the whole picture of the proposed model, and introduce each important part.

(v)              Since the Mish activation function was not proposed by the authors and the authors did not prove the effectiveness of Mish mathematically, From the functions presented in Figure 7, we can see that SiLU and Mish have minor differences. I am not sure what the authors intend to prove in Section 3.1. If no new knowledge was contributed, then please shorten this section by citing references. Besides, CBM was already introduced in YOLOv4, it is also not new to the field.

(vi)            In Section 3.2, the content did not fit. As the title suggested, D2F, however, the authors mainly described C2F, and the technical details of D2F were unclear. How were the input images split? What was the architecture of the bottleneck? What were the shapes of feature maps? What did ‘C’ mean? What was the output shape?

(vii)           In Figure 13, it is also necessary to present the validation result.

(viii)          In the experiment section, readers may also be interested in the result of model parameters and detection speed. Please add these two criteria in Table 2.

(ix)             Throughout the experiment section, the authors have to conduct an in-depth analysis of their experimental results, not only describing the value changes, especially for Figures 15, 16, 17, and 18.

(x)              There lacks a discussion section. The authors have to discuss the limitations of their detection methods, as well as to compare their work with others.

 

(xi)             The conclusion section did not cover the main content of the manuscript. The authors must re-organize it. 

Extensive English editing is needed.

Author Response

Thank you for allowing a resubmission for our manuscript, with an opportunity to address the reviewers’ comments.

We would like to express our gratitude for granting us the opportunity to resubmit our manuscript and address the valuable comments provided by the reviewers. We have attached the following documents.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The Ms has been improved 

Minor revision for English is needed 

Reviewer 2 Report

The authors have successfully addressed my previous concerns and I have no further comments.

Moderate Englishe editing is needed.