A Systematic Overview of Recent Methods for Non-Contact Chronic Wound Analysis
Round 1
Reviewer 1 Report
Comments to the Authors
The paper realizes “a systematic overview of recent methods for non-contact chronic wound analysis”. After an exhaustive revision of this paper, I consider that the work presents a remarkable effort of compilation and synthesis in a wide field of study like wound analysis. However, from my point of view, this work, although significant, is at the moment incomplete. For this main reason, and other minor aspects, my recommendation is that this paper would be reconsidered after major revision.
Main aspects:
In my opinion, a review paper not only should compile the complete catalogue of knowledge, but also demonstrate “what is known and what is unknown” [1] in this field of study: non-contact chronic wound analysis. In this case the paper seems to be a compilation of more than ninety articles and other works classified and sorted by pre- and postprocessing methods and main wound imaging tasks.
I think the paper, at this moment, is only a valuable work of compilation that presents shortcomings such as follows:
- A good review paper should advance from history to final synthesis, through a brief explanation and classification of methods and an analysis of main applications and results. Above all, the fundamental objective should be to determine “what is known and what is unknown” [1], determining advances in future and new research lines. The authors should be highlighting this question in more detail. I suggest that the authors read the editorial of Allwood and Erman Tekkaya cited below.
- A reader that revises this paper should know answer this question: where is knowledge complete and where are the gaps in non-contact chronic wound analysis? [1] A partial explanation is developed in Conclusion section but it should be expanded.
- There is not any comment from research methodology used in this work. What is the body of knowledge? What are the sources of information? What databases were employed? How the papers were selected and sorted (number of references, year of publication, relevance of journal, etc…)? The authors do not propose a clear taxonomy for classification and validation. They only refer the aforementioned classification by methods and wound imaging tasks but this is not enough.
- The main wound imaging tasks seem to be extracted from a previous review conducted for Zahia et al. in 2020 [2] in Artificial Intelligence in Medicine journal. I recommend to authors a complete revision of whole Section 2 of this paper. In this section the authors will find the answers to my previous comments.
- The Discussion section is missing in the paper. For what reason did the authors not include this section? Shouldn't a review article have an effective Discussion section? I think the present Conclusion section has paragraphs and contents that should be exposed in a Discussion section.
I suggest a deep revision of paper. At the moment the paper seems to be a report that contains a catalogue of papers with a brief description and, finally, some general conclusions. All the sections should be revised and rearranged. The classification must be more justified and supported by well-proven arguments and taxonomy. For this main reason a research methodology should be proposed and validated. All the conclusions and references should be checked according this methodology. In addition, it is necessary a new Discussion section. A review paper is a hard work but a great value to other researchers for this, finally, I want to reaffirm the potential value of work and I encourage to authors to continue with it.
Other minor points:
- Generally, the references do not maintain a common structure and some of them contain errors (K & B [13]?; Kumar, K. S., & Reddy, B. E.? / Reference 1, available online?). The authors should revise all references and their text format. I recommend to authors revise the journal reference style.
- In Section 2 the first paragraph is referred to reference 19; Are there other performance metrics? Are the introduced terms a contribution of the cited paper? Why is reference 19 in uppercase?
- In Page 6, the CIELUV method needs a better explanation.
References:
[1] Allwood, J. M., and A. Erman Tekkaya. Writing a review paper. Journal of Materials Processing Technology, 2012; 212, 1-2. doi:10.1016/j.jmatprotec.2011.09.003
[2] Zahia, S.; Garcia Zapirain, M. B.; Sevillano, X.; González, A.; Kim, P. J.; Elmaghraby, A. Pressure injury image analysis with machine learning techniques: A systematic review on previous and possible future methods. Artificial Intelligence in Medicine, 2020; 102, 101742.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
1. The document addresses comprehensively the literature on the topic of methods for non-contact chronic wound analysis, in a very structured way. It starts describing the main characteristics of these wounds and continues with the description of the performance metrics and algoritms. After that, it deals in sequence with the stages of: a)image acquisition, preprocessing and postprocessing, b) wound segmentation, c) tissue classification, and d) 3D wound reconstruction.
2. Some review of English grammar and puctuation in the document would be necessary in order to improve its readability.
3. A short historical analysis of the advances in the field of study -a timeline maybe- would help to put the topic into a proper context.
5. Any research paper should specifically include a clear and detailed explanation of the methodology used for its development.
6. Given the importance of such matters, the challenges to the topic of analysis that are pointed at in the first paragraph of page 3 might require a more detailed explanation, both from the perspective of the wound -type, location, combination, interference with other skin conditions, skin tone, etc.), and of the aquisition techniques (camera type, resolution, orientation, lighting, distance, lens focal length, etc.).
7. In the formulae at pages 4 and 6, the final comma/apostrophe results somehow confusing; eliminating or displacing them some spaces to the right would be of help.
8. In the table at page 9, the last cell of the first row -after the title- shows '5min' at the end. A similar thing is observed in the last cell of the 16th row of the table at page 13. Do these refer to minutes? It might be more convenient to keep coherence of units by using seconds. Some rows in the same table show what appear to be blank lines -for example, the 5th, 7th, 11th and 16th; is some text missing?
9. A summary table in section 6 would help to maintain coherence with the structure of sections 4, 5 and 7.
10. A 'Discussion' section is missing in the paper, where its core findings are put in context and evaluated, while at the same time the field areas were insufficiently addressed are explained, in order to help the development of the state of the art.
11. While the 'Conclusion' section summarizes quite conveniently the comparative analysis made on the literature studied, it doesn't address in a comparable depth the field areas where further development is required in order to extend the capabilities and functionality of the methods for non-contact chronic would analysis.
12. An 'Acronyms' section would help to clarify many of these terms that are used profusely along the document: MOWA, WITA, RGB-D, YCbCr, YDbDr, etc.
I look forward to read the finished version of the manuscript with the appropriate modifications, as it addresses a topic that is of much interest to me.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors provide a overview of papers which are dealing with all types of wounds in the field of the wound image analysis. My recommendation for the article is "Major Revision" because of the following reasons and comments:
A discussion and analysis section is necessary.
Paragraphs are necessary with lines of future work that are deduced after this review.
Some scheme is needed to help follow the paper.
In relation to the sentence: ‘Knowledge and fast decision making are crucial at the beginning of the wound treatment because if wounds are not accurately and orderly treated, the severity of the wound can increase drastically’, the decision-making process should not only be crucial in early stages, but also on an ongoing basis during the wound healing process itself.
In relation to Figure 1, which shows the types of tissues, is it your own image or was it provided by a third party for this review?
In relation to the sentence: ‘The only available wound image database is Medetec [17], which is widely used by researchers in this field [11, 18, 6, 4]’, it is considered that the authors should not be categorical. There may be other databases that they have not found.
The authors comment that: ‘Algorithms for CW image analysis can be divided as supervised and unsupervised’, however, I would like you to clarify whether approaches that also incorporate reinforcement learning could be considered or exist in the current literature.
I am missing a paragraph in the introduction section and in the conclusions to the Clinical Decision Support Systems, since in a certain way these tools could be integrated into other higher order systems, providing value information.
Page 1, …toward healing in 30 days… > Final quotation marks are missing in the sentence
Page 6 > Use a different bullet depending on the indentation level.
P7 “authors in [40]...”, P10 “Authors in Maity et al. [9]…”, P17 “Authors in [51]…” “In [9]…” Research like [29]…” > Avoid these expressions as much as possible.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
From my opinion the paper has been clearly improved. All my previous comments have been answered adequately by the authors. So, my proposal is that the paper be accepted for publication in the Applied Sciences journal.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
1. The manuscript has been modified along the lines suggested in the previous comments, and now it mostly addresses the issues that were indicated in that review.
2. I would advise some more review of English grammar throughout the document, especially in the new incorporations, and also to pay attention to the punctuation. A language review by a native speaker could be of help.
3. Font size in the text of Figure 2 is larger than the regular text in the document. I would recommend reducing, either font size or figure size to keep the balance between both texts.
4. In chapter 9 'Conclusion', the claims made in the two larger paragraphs correspond to the Author's views, I understand. If that is the case, such circumstance should be indicated somehow. If that is not the case, the origin of such claims must be expressed instead.
5. I'm looking forward to read the finished manuscript after its publication, because of its interest for my field of research.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The authors have made the corrections and revisions that I have suggested. After this review, the article has improved considerably. Therefore, my recommendation is to accept the paper for publication.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf