Research on Contactless Detection of Sow Backfat Thickness Based on Segmented Images with Feature Visualization
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsI have one comment on the work: there are too few teaching cases. Therefore, I suggest adding in the title that we are talking about preliminary studies.
The work is well organized, correct in terms of the methodology adopted. The whole work is written well and clearly. The reader can assimilate the knowledge contained in the work in an accessible and comprehensive way. However, as I mentioned earlier, please emphasize in the text that this is preliminary research. The group is unfortunately not representative. This does not detract from the work, but it should be noted.
Additional comments:
It is very rare to find such a methodically good work. Really the only complaint is the very small number of teaching cases.
1. What is the main question addressed by the research?
The goal of the study was defined in detail and clearly.
2. Do you consider the topic original or relevant in the field? Does it
address a specific gap in the field?
The topic is interesting, not groundbreaking, not innovative. In my opinion, it brings new scientific knowledge.
3. What does it add to the subject area compared with other published
material?
Works and systems related to animal evaluation are scarce. This is primarily due to the technical difficulties in acquiring an image of an object and describing it properly. Life evaluation of an animal and identification of valuable elements is very difficult - the authors prove that it is possible. In my opinion, there should be more teaching cases.
4. What specific improvements should the authors consider regarding the
methodology? What further controls should be considered?
The authors adopted the correct methodology and applied it correctly. The only improvement is to obtain more research material and produce models on a larger sample.
5. Are the conclusions consistent with the evidence and arguments presented and do they address the main question posed?
Yes
6. Are the references appropriate?
Yes
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsArticle: “Research on contactless detection of sow backfat thickness based on segmented images with feature visualization”
Observations:
1. In the introduction it is mentioned “Aiming at the problem that the existing methods for detecting sow back fat thickness are 13 stressful, costly and cannot be detected in real time, this paper proposes a non-contact detection 14 method for sows with residual network back fat based on image segmentation based on feature 15 visualization of neural networks”. Derived from this, I have the following observation. A) It is important to explain what the main contribution is; it is only the application, it is important to detail the novelty of this work. B) Mention that the state-of-the-art method cannot be detected in real time, why in the proposed method if it is possible, add more details about it. C) The authors should add a comparative analysis against the state of the art to strengthen the results and conclusions.
2. Line 19, what does the unit of measurement (mm) represent?
3. Is important a hard review of the template of the Applied Sciences. 3. It is important to carry out a thorough review of the Applied Sciences template. This document does not use the correct format. There are many inconsistencies when referring to Figures, the authors use Fig. and Figures, also the references are not cited correctly and in some of them there is no adequate citation (line 62, 83 Fernandes et al.). Review the complete document.
4. Some words use capital letters where they don't belong (you can find some on lines 60, 66, 99, 134, 175,259, 383). Review this this in all the paper.
5. Line 209, 211. The name of the authors is redundant.
6. Avoid using these types of contractions (sow's back, pig's body, network's decision-making, cow's rump). Review all the paper.
7. Line 237, redundant use of i.e.
8. There are many typos, it is important to review the entire paper and improve it.
9. I think the data set is too small and maybe the author needs more data to get better validation results.
10. Provide more details on how the architecture presented in Figure 2 was selected. Why this architecture?
11. Line 173, add reference about OPTUNA.
12. Line 259. What figure is it referring to?
13. Review the format of references, correct and add more references related to the work.
Comments for author File: Comments.pdf
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThis paper presents image-based analysis for measuring sow backfat thickness, as a contactless method to reduce stress and effort in the process.
Please, find my concerns bellow:
The abstract seems lengthy. It also refers to a P2 measurement point, which as not been yet contextualized. I propose to change that P2 point to some other term more intuitive at this point of the paper.
The introduction of the paper starts strangely: “The country is a major pork consumer and farmer, and as the population grows, the demand for pork is increasing The demand for pork is increasing with population growth”. Which country? Don’t forget you are writing for the scientific community worldwide; therefore, it is better to clearly mention the country you are referring to.
The authors have parts of sentences that appear repeated: “China is not a strong farming country, and there are many problems in China's hog farming, such as high cost, low degree of automation, and weak breeding capacity of sows[1]. However, China is not a strong farming country…”. Also, this sentence may raise the sense that the authors are contradicting themselves, if we compare to what is said in the first sentence: “The country is a major pork consumer and farmer…”. These kind of “phenoms” must be banned from the form like the authors expose the content. In that sense, the paper must be thoroughly revised to before publication.
The issue of backfat thickness measurements appear to be addressed before (e.g .works [5] and [12]), but it is not clear what this paper brings of “new”, comparatively to previous work.
The dataset construction seems to be clearly explained. In L133-134, the last sentence seems to be incomplete.
In the results analysis, other metrics could have been used, such as IoU, Dice/F1-score, Precision, Recall, etc. These ones are extensively used in the literature and favour results comparison.
Finally, this paper is missing a discussion comparing these results with the previous proposed works done in the field with similar purposes.
Comments on the Quality of English LanguageEnglish needs to be extensively revised, mostly because of the repetitions and "contradictions" mentioned in the comments/suggestions for authors.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors improved the paper and made the changes indicated in the previous revision. I think the writing format just needs to be improved.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors addressed most of the proposed aspects highlighted in the first round.
Although, the discussion is still poor. The only sentence added (at least, highlighted by themselves) in this context, was the following:
"The results showed that the estimated and measured backfat thickness of sows based on the residual network model of the present study had a good linear relationship, the model prediction accuracy was high, and the experimental results of this paper are better than those of the five groups of Yu Mengyuan et al[27]."
Only a single comparisions with other works developed for similar goals was very superficially done, considering a previous paper of the same authors. Although, others could have been considered. For example:
https://www.ingentaconnect.com/content/tcsae/tcsae/2015/00000031/00000018/art00035
https://www.sciencedirect.com/science/article/abs/pii/S1537511017308590
or eventually this one: https://www.sciencedirect.com/science/article/abs/pii/S0168169923006579
Regarding the last one, that uses ultrasound, it would be very interesting to have a discussion around the diferences of the proposed methods and instruments, some results comparision to highlight the most accurate approach, and some notes formulating hypothetical reasons for those differences, as well as, for example, identify which is the less expansive setup for farmers...
Therefore, this paper requires several improvements in the discussion section, which is an essential part of every paper that is submitted to a JCR journal like Applied Sciences.
The English seems better, now.
Author Response
Please see the attachment
Author Response File: Author Response.docx
