Segmentation of Acne Vulgaris Images Techniques: A Comparative and Technical Study
Round 1
Reviewer 1 Report
The authors present a review/comparative study of segmentation of acne vulgaris image techniques.
The review of existing literature is comprehensive, and the content is in general of interest.
The main criticism of this work is the very poor standard of proof checking demonstrated. There are numerous formatting and typographical errors throughout the document that are frankly unacceptable. It is clear that this document did not undergo sufficient peer review before submission. This will have to be addressed before resubmission.
The paper does a reasonable job of summarizing the state of the art, and highlights the difficulties in comparing acne segmentation methodologies. However, the conclusion is too brief. Expand, indicating the actions/steps necessary to address this.
Specific comments:
Page 1 - Author list: Surname of last author is missing
Page 1 - Introduction: citation for Ramli et al contains the full paper title. This error is repeated throughout the document. Ensure citations match the required theme
Page 2 - First paragraph: incomplete sentence: "...new computer-based techniques"
Page 2 - Table 1 is misaligned
Page 2 - Sentence "Studies not published in English" has no context. This appears to be an internal note not removed before submission
Page 3 - Paragraph 2-3: running sentence is split between two paragraphs
Page 3 - Section 3.1 There is a blank line between the title and the text. Remove.
This issue is repeated multiple times in the document. Review and fix all instances
Page 4: inappropriate use of hypen ( - ) " - in this case chi-squared test was calculated - ". Recommend replacing "-" with round brackets. Note there are multiple instances of this through the document
Page 5-6: Table 2 is split across two pages. Ensure table fits into a single page
Page 5-6: formatting of table - the "limitations" column is too narrow for the amount of text. Reformat
Page 9-10: table is misaligned with the text. Also, the format of table 2 (text, borders) does not match the theme in the rest of the document
Page 13: Discussion section - break text into paragraphs. Fix typos (multiple instances where there is no space after brackets.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
The article “Segmentation of acne vulgaris images techniques: a comparative and technical study” by Moncho-Santonja et al. is devoted to reviewing the current state of image segmentation techniques for acne vulgaris.
The article can be of interest to a large scientific and clinical audience. However, in its current form, it has multiple shortcomings:
Major issues:
1. In its current form, it is a narrative literature review. It lacks multiple characteristics of the systematic review (see my comments on methodology). The methodology questions need to be addressed. Alternatively, it can be reclassified into a narrative literature review or scoping review.
2. Honestly, I have not understood how the authors segmented studies into two groups. From my understanding, there are two approaches: a) standard statistical methods (including K-means) with or without color transformation, and b) “black box” deep neural networks, which typically do not require color transformations. Most techniques, which the authors mentioned in the 3.2 section (e.g., K-means and SVM, which are statistical models, although used in ML), can be considered “classical” techniques. Alternatively, they can be split into 3 groups: simple methods (thresholding), statistical methods, and deep learning. But combining K-means and CNN in one bucket does not make sense.
3. Methodology:
· What was the query used for the literature search?
· What were the research questions?
· Was the review protocol registered with PROSPERO?
· Critical Appraisal (Risk of Bias Assessment) is missing, which is mandatory for the systematic review.
· The level of evidence is not assessed.
Other issues:
4. As this is a systematic review, including the flow chart with the overall number of retrieved publications and eliminated publications at each step is recommended.
5. The authors use a reference style, which is not typical for this journal
6. Page 5, 6th sentence from the top: Does (Abu Zaki, Ibrahim, & Sari, 2019) study qualify as fluorescence-based? They use smartphone LED
7. Tables 2 and 3 look unprofessional. Consider turning it into landscape mode.
8. On pages 6 and 8, Table 2 has additional sets of column titles. Are they new tables? The same issue with table 3
Some minor shortcomings (it is not an inclusive list. There too many of them):
9. Page 2, 2nd paragraph from the top: The sentence is broken down into 2 parts by a line break. Also, in this sentence, there is strange punctuation using dashes. The authors use this unconventional punctuation on several occasions. Consider changing it to more standard using commas.
10. Table 1 looks odd. Consider proper formatting
11. Page 2, 2nd paragraph from the bottom: Consider turning exclusion criteria into a bullet list.
12. Page 3, 2nd paragraph from the top: An unjustified line break after the word “machine”
13. Page 4, 2nd paragraph from the top: The first sentence looks odd “Ramli et al. work in (Ramli, AS, & AFM, Identification of acne lesions, scars and normal skin for acne vulgaris cases., 2011) with CIELAB color space”. There are multiple issues with this sentence. 1) Unusual citation “(Ramli, AS, & AFM, Identification of acne lesions, scars and normal skin for acne vulgaris cases., 2011”. The same in the next paragraph. 2) Reference should immediately follow Ramli et al. The word “work in” is inappropriate. The same issue can be found in many other instances
14. Page 5, 4th sentence from the top: “imagens”?
15. Page 5, 5th sentence from the top: For an imaging system VISIA-CR the manufacturer and county are required.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Response to reviewer feedback: the authors response to comments regarding typographical errors, formatting etc was in general "this is not the final version". While this may be, the quality of the submission was regardless not of an acceptable standard. Regardless of whether this submission is the final format are not, the authors are strongly encouraged to acknowledge that the quality of the first submission was not of sufficient standard, and that future submissions to journals will not be of this low standard, as a courtesy to the journal editors.
Section 2 - methods: excluding non-open access papers is troubling, this excludes a very large section of the available literature and undermines the validity and usefulness of this review. This needs to be addressed before this review can be considered for publication.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 2 Report
The article has been significantly improved during this round of revisions. Therefore, it can be published if several deficiencies are addressed.
1. Several methods seem misclassified:
a) Page 7, 2nd paragraph: (Malik, Humayun, & N, Novel Techniques for enhancement and segmentation of acne vulgaris lesions, 2014) It seems that authors describe here K-means, so according to their classification, it is ML – other ML Methods.
b) Page 7, last paragraph: The method described in bullet points (Min, Kong, & Yoon, 2013) is clearly an ML-based method (namely, using ANN)
2. Figure 1: The added figure is very helpful. However, it provides a lot of confusion:
a) There is a discrepancy between the number of studies included (27) vs. the number of studies mentioned in the abstract (29),
b) The flowchart itself is confusing: the number of articles that come through arrows from a box should add up. For example, the number of "full-text articles assessed for eligibility" should be 41, which is further split into "excluded with the reason (n=14) and "included in synthesis (n=27)". The whole flowchart needs to be reworked.
Minor deficiencies:
3. Page 3, 2nd paragraph: There is a contradiction in the following statements:" No exclusion criteria related to publication date were applied. The exclusion criteria were the following:"
4. Page 3, bullet list: "amplified images." Did the authors mean "magnified"? Amplified is an ambiguous term in this context.
5. Page 7, 2nd paragraph: it is unclear what the authors meant by "Mapping of RGB image pixels to visible spectrum is proposed, as a way of visualizing little color changes (spectrum 380-720 nm range is obtained with a band separation of 1 nm)." Is it hyperspectral imaging? Otherwise, I can't understand how you can map 3 channels into bands with a 1nm width. Consider clarification.
6. There are several incorrectly placed hyphens, e.g., "heal- thy" (Page 7, 2nd paragraph)
7. There are several sentences with grammar broken with references. For example, the sentence "In the study developed in (Malik, Humayun, & N, Novel Techniques for enhancement and segmentation of acne vulgaris lesions, 2014) opt to increase dynamic range on YCbCr images in order to augment contrast and enhance lesions visualization" (page 7, 2nd paragraph) seems odd (no subject?). Proper grammar is required. There are several similar instances across the text.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 3
Reviewer 1 Report
My primary concerns relating to this work have now been addressed. There are some minor editorial issues that should be addressed, listed below. Assuming these issues are corrected, I would recommend publication.
Page 2 - Methods: exclusion criteria should be listed as bullet points to improve clarity
Page 4, line one - reference to Jena et al is in bold font. This should be standard text font.
Page 5, start of second paragraph - replace "There is also" with "There are also", and "that develops" with "that develop"
Page 4, section 3.1, paragraph 3 - the fond for the references is different than the main body of text, please harmonize
Page 8 - replace references to "Mathworks" code with "Matlab". Matlab is the software, Mathworks is the company that created Matlab.
Page 8, last paragraph - reference "hasanah et al", capitalize "Hasanah". This error is made throughout the text.
Author Response
All the minor changes have been modified.
Reviewer 2 Report
The manuscript has been improved in the latest round of revisions.
The manuscript can be published if the following minor concerns are addressed:
1. The arrow between blocks a and b in Figure 1 looks odd. Please replace it with the same type of one-segment arrow as all others.
2. Exclusion criteria “Those publications which are neither open access nor available to Polytechnic University of Valencia members” can be a significant shortcoming. Does it mean that if the abstract was available, but the full text was not, then the article was discarded? It is unclear and needs to be clearly explained. With the proliferation of open access, I don’t consider it a reason for rejection. However, it needs to be clearly stated and discussed as the weaknesses of the study in the Discussion.
Also, it should be noted that figure captions and references are not in the journal format. While the authors mentioned that the Editor permitted it and they will be replaced in the final version, I consider it strange.
Author Response
- The arrows in Figure 1 have been modified.
- The exclusion criteria has been explained in more detail by adding: Only six of the 100 records excluded after screening by title and abstract met the exclusion criteria number three. Special attention was paid to these records and also no segmentation method used was described in the abstract. If those publications were considered relevant the Polytechnic University of Valencia has a special service for specific queries to studies with restricted access. It has also been considered and added in the Discussion section by adding: Related to the design of this study, exclusion criteria number three could be considered its main weakness. As shown in Figure 1 all the records (title and abstract) found were included (Identification step). In the Screening step, we identified those for which we did not have access to the complete work and had to discard them. However, they were not discarded directly; the abstract was analyzed and, since they were not considered relevant, the exclusion criterion was applied. As explained above, if the records would have been considered relevant, an institutionally special access to the full text could have been requested.
- The figure captions and references have been adapted to the journal format.
