YOLO-Based Model for Automatic Detection of Broiler Pathological Phenomena through Visual and Thermal Images in Intensive Poultry Houses

Round 1

Reviewer 1 Report (Previous Reviewer 2)

There was still a need for some revisions. I reported these fixes. Attached you will find this file.

Overall Recommendation : Accept after minor revision

 

 

 

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report (New Reviewer)

1. There are several grammar and syntax errors throughout the manuscript. Proofreading is recommended to avoid any errors in the entire manuscript, especially in the abstract.

2. The authors should clarify two types of models since both are thermal-based and visual-based models. What is the key difference between these two models? What is their main methodology and what are they using for?

3. I wonder how the environmental conditions such as temperature and humidity inside the poultry house affect the efficiency of the models, especially the thermal-based model?

 

4. What is the range of temperature that the thermal model can achieve the highest accuracy?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report (New Reviewer)

1. Repeat in line 25 "Correspondence:"

2. Figure 3A-3D in lines 229-236 is inconsistent with Figure 3A-3D in Figure 3.

3. It can be seen from Figures 1 and 3 of the paper that the breeding density of chickens is relatively low, while the actual breeding density of chickens is relatively high. Will this affect the final test result?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 4 Report (New Reviewer)

In this study, using a YOLO-based model, the authors used visual and thermal images to detect pathological phenomena in broiler chickens. The outcomes of this study have an excellent effect on poultry production; however, the writing of the manuscript is very poor. They need to focus more on writing the manuscript to improve its quality. Please find my comments below:

General Comments

·       This manuscript is overly detailed. The authors used a lot of information in all the sections of their manuscript. Please present your data in a concise and precise form. You need to minimize the volume of your manuscript. Please check and correct it throughout the manuscript.

·       This manuscript has space and punctuation issues. Please correct them throughout the manuscript.

·       Please use “thermal- and visual-based” instead of “thermal-based and visual-based” throughout the manuscript.

·       Abstract should be improved, especially the introduction and conclusion. The introduction of your abstract is too long, and you didn’t mention the conclusion of your abstract. Also, the methodology and results sections aren’t up to the mark. Please improve your abstract.

·       The introduction of your manuscript must be improved. You provided a lot of information. This section seems like a review and literature of a thesis. Please minimize the volume of this section. The introduction needs to be direct, specific, and clear. The text must be readable and as short as possible.

·       Please provide the company or manufacturer’s name (with their city and country) of any instruments you used in your study.

·       Please check the present and past forms of your manuscript. You used the present form in several sentences when you should use the past form. Please correct them throughout the manuscript.

Abstract

Line 26-29: Please minimize the introduction of the abstract. Try to keep it in a sentence.

Lines 38 and 41: Both models are thermal- and visual-based, but you’re saying first and second type. Please clarify this.

Line 62-64: Please check this sentence. It seems incomplete. Please correct it if necessary.

Line 69-89: Please minimize the volume of these paragraphs. Use only 1-2/2-3 sentences for each feature.

Table 1: Please omit this table. Just cite those articles which are suitable for the manuscript.

Line 141: “This paper” should be “This study”.

Line 162: “aims” should be “aimed”

Line 163-165: Please keep these under the objectives of your study. This seems like a conclusion of your study. Please write it properly.

- Your manuscript's materials and methods section looks like a results section. Please mention those methodologies you used in your study in a concise and precise form. You don’t need to provide all the information.

Figure 1: You don’t need to use this one. Please omit it

Line 168-193: Please provide these characteristics in a short form. If you need any references, please use them.

 

- You didn’t mention and describe Figure 9 in your manuscript. Please check it.

The manuscript writing is very poor. The authors must improve the manuscript by minimizing its volume. They used a lot of information like they were writing a thesis. I believe they need to check this issue properly.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 4 Report (New Reviewer)

The authors addressed all of my comments adequately. However, if it is possible to minimize the volume of the introduction section, please try to do it.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Overall, this manuscript intends to use Yolo-based object detection techniques to classify the pathological phenomena of broilers in thermal and visual images. It is of reference value to use thermal images to classify pathological phenomena to reduce the influence of light conditions on recognition. But the focus of the research in this manuscript is not clear. The experimental results do not have strong logic with each other. In addition, which pathological phenomena are labeled and classified in this study, the authors should revise the manuscript and maintain consistency. The following concerns also should be addressed.

Abstract:

1.       Line 32-34: Does this mean there are an average of 5 broilers on one image? It does not match the intensive poultry houses in the title and keywords.

2.       Line 35-38: “two different object detection models” “The first model” “The other model” seems vague. Could you please explain them better?

3.       Line 40-41: This sentence needs to be more specific. Which model?  “The first model” or “The other model”.

Introduction:

1.       Line 61-63: Which feature was used in this research? Its advantages over other methods should be mentioned below.

2.       Line 141-143: This sentence seems contradictory.

3.       Line 153-156: This table needs to be more informative and persuasive. “Unsuitable" is not reflected in the table.

4.       Line 157-158: please review this sentence.

Materials and Methods:

1.       Line 168-169: Will the thermal image acquisition be affected by the distance between the camera and the broiler?

2.       Line 169-171: “the accuracy is ±2.0 °C or ±2%” There is generally little difference in body temperature between broilers, have you taken any additional steps to ensure the accuracy of thermal image acquisition? Especially for “diseased eyes”.

3.       Line 180-182: This sentence needs to be more specific. Is "unhealthy" also a kind of label? In addition, it is possible that these behaviors appear in a broiler at the same time, and how this situation is labeled.

4.       The resolution of Figure 3 needs to be improved so that the reader can see the text clearly.

5.       Line 222: Could you elaborate on the processing and annotating of thermal images? It is better to use diagrams, such as flow charts.

6.       Line 246: The hyper-parameters of YOLOv8 should be added to Table 2

Results and Discussions

1.       Line 432-433: Yolov7 and Yolov8 differ in network structure and technology, and the conclusions drawn by experiments conducted by different versions of Yolo are not scientific and persuasive.

2.       Line 459-461: Visual image dataset should be included in the title of Table 4.

3.       Line 504-505: I don't think it makes sense to compare the impact of augmentation methods on model performance in detail, which should be briefly described in ‘Materials and Methods’.

4.       Line 512-513: Could you explain in detail how optical images and thermal images are combined, and the manuscript seems to train on two sets of datasets separately.

5.       Line 525: "Low activity and appetite" refers to "lethargic chicken" above? The term should be consistent across the text, please review the whole manuscript

6.       Could you explain why the visual image dataset mostly performs better than the thermal image dataset in Table 5, but in Table 4 thermal image dataset performs better than the visual image dataset?

7.       Line 529-531: The resolution of Figure 13 needs to be improved, and the detection results of every pathological phenomenon should be presented.

Conclusions

1.       Line 541-542: Why not include the “pendulous crop” mentioned in the abstract?

2.       Line 544-545: Which version is the best one? The results of the comparison should be given.

3.       Line 554-555: What are these metrics used to evaluate? Broiler detections or the pathological phenomena classification?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

Dear Author(s),

I think it will be a better-quality publication if the revisions I have mentioned in the attachment are made.

With all respect.

Comments for author File: Comments.pdf

Author Response

Please see the attachment

Author Response File: Author Response.docx