The Potential of AI-Powered Face Enhancement Technologies in Face-Driven Orthodontic Treatment Planning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is an interesting article examining the attractiveness of faces modified by AI technologies.

Despite the idea being interesting, the manuscript needs some improvement. 

- the title is not appropriate. The research does not focus on orthodontic treatment planning. The fact that the results of the present study could potentially be used for orthodontic treatment planning is, based on the present paper, only a possibility.

- lines 38-41 are not scientific or appropriate.

- lines 47-50 are not needed.

- explain AI abbreviation the first time, and afterward only use AI in the abbreviated form.

- lines 147-148 are not scientific or appropriate.

- How was the sample size decided?

- only 5 pairs of pictures were analyzed for anthropometric point. how was the number decided? Please consider examining all pictures.

- how many people were invited to revise the pictures?

Author Response

This is an interesting article examining the attractiveness of faces modified by AI technologies.

Despite the idea being interesting, the manuscript needs some improvement. 

Response: Dear reviewer,

Thank you very much for your insightful comments. We tried to incorporate your comments and ideas, and we feel they have improved the quality of our paper.

- the title is not appropriate. The research does not focus on orthodontic treatment planning. The fact that the results of the present study could potentially be used for orthodontic treatment planning is, based on the present paper, only a possibility.

Response: Regarding the title of the article, the whole article points towards the use of Artificial Intelligence (AI) and FaceApp (or similar applications) in orthodontics and we believed that this title would suit the purpose of the article. On top of that, being orthodontists, clinical application was the direction we wanted to pursue.

- lines 38-41 are not scientific or appropriate.

Response: Lines 38-41 were rephrased so that they emphasize the need for patient-oriented treatment planning.

- lines 47-50 are not needed.

Response: Lines 47-50 are quite important when it comes to the aesthetic aspect of orthodontics. We need to appreciate the need for both stable occlusion and smile aesthetics and facial harmony.

- explain AI abbreviation the first time, and afterward only use AI in the abbreviated form.

Response: Thank you, AI is explained the first time it is used.

- lines 147-148 are not scientific or appropriate.

Response: Lines 147-148 were rephrased.

- How was the sample size decided?

Response: We decided to have 100 faces in total based on a discussion with our statistician. For chosen statistical methods used to analyze the dataset, this number of objects was sufficient.

- only 5 pairs of pictures were analyzed for anthropometric point. how was the number decided? Please consider examining all pictures.

Response: The part of the study which analysed anthropometric points and distances, aims to correlate given face changes with the biggest attractiveness score changes. If we analyzed changes in all pictures and put them together, it would be less obvious which changes might have pointed towards the biggest attractiveness shift. For our upcoming paper we are currently developing a computer programme for automated analysis of facial landmarks. If successful, we will surely run all faces through this programme, and perhaps we would get some novel findings.

- how many people were invited to revise the pictures?

Response: Each questionnaire was sent to 1800 respondents in total. 

We very much appreciate your effort to help us improve this article. Based on your constructive comments, we explained the study design in greater detail and strengthened the discussion. Thank you again for taking your time to thoroughly review our submitted paper.

Kind regards.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

the article is novel and presents a new perspective. However, you still have to apply some changes to it:

1. In the introduction, please write more about the facial / general aesethetics, basing on Fibonacci's perspectives - please, note the importance of perception of symmetry (facial and dental) and smile esthetics and the golden ratio. Also, add the information on how important is dental symmetry in overall perspective (eg. ttps://doi.org/10.3390/sym14061257)

2. Line 97, give the reference number just after the name - this should refer to all citations like that

3. Line 159, give the age of the individuals (mean + SD)

4. In the discussion, add the information what may influence the decision for orthodontic treatment (eg. doi: 10.17219/dmp/151577)

Thank you

Author Response

Dear Authors,

the article is novel and presents a new perspective. However, you still have to apply some changes to it:

Dear reviewer,

thank you for your comments and suggestions.

Comment: 1. In the introduction, please write more about the facial / general aesethetics, basing on Fibonacci's perspectives - please, note the importance of perception of symmetry (facial and dental) and smile esthetics and the golden ratio. Also, add the information on how important is dental symmetry in overall perspective (eg. ttps://doi.org/10.3390/sym14061257)

Reply: 1. Based on your suggestion, we added more text on facial aesthetics. Thank you for the reference that you provided. It was an interesting article to read.

Comment: 2. Line 97, give the reference number just after the name - this should refer to all citations like that

 

Reply: 2. The references were adjusted.

 

Comment: 3. Line 159, give the age of the individuals (mean + SD)

 

Reply: 3. The respondents did not state their exact age; they only ticked the age category. Based on a consultation with a psychologist, some people do not feel comfortable stating their age even in anonymous polls. Identifying oneself with age groups is considered more acceptable.

 

Comment: 4. In the discussion, add the information what may influence the decision for orthodontic treatment (eg. doi: 10.17219/dmp/151577)

Reply: 4. The main rationale for undergoing an orthodontic treatment was added to introduction

 

Thank you

Kind regards.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The paper presents an innovative and timely exploration of how AI-powered face enhancement technologies, specifically FaceApp, could potentially be applied in orthodontic treatment planning. The content is highly relevant to current trends in both AI applications and orthodontics. The study addresses an important topic with significant potential to contribute to the field, demonstrating creativity in its approach and methodology.

Suggested improvements:

1. Provide a more explicit statement of how the findings might translate to real-world orthodontic practice.
2. Further analysis into the nature of specific facial changes, which caused an increase in the attractiveness scores.
3. Discuss more critically how biases in AI-enhanced photographs would influence orthodontic planning and treatment.
4. Provide additional references for claims about AI capabilities in the introduction.
5. More in-depth detail about the specific FaceApp algorithms/filters used to better reproduce them.
6. Clarify why you analyze in detail only the five pairs with the greatest attractiveness difference.
7. Give some measure of interrater reliability of the attractiveness ratings.
8. Move some information from the discussion (e.g., about FaceApp's algorithms) to the methods section.
9. Ensure the consistent use of first-person pronouns throughout the paper. 
10. More citations should be added to some statements found in the introduction and discussion sections.

In conclusion, while the paper is highly promising and treats a critical topic, it needs moderate revisions in order to meet publication standards. Focusing on enhancements to the discussion section, clinical implications, and potential biases, organization of results, and improved writing, the authors could realize major improvements in the impact and clarity of their work. These improvements are what will make this innovative research find its value addition to the current knowledge on AI-assisted orthodontics.

Author Response

The paper presents an innovative and timely exploration of how AI-powered face enhancement technologies, specifically FaceApp, could potentially be applied in orthodontic treatment planning. The content is highly relevant to current trends in both AI applications and orthodontics. The study addresses an important topic with significant potential to contribute to the field, demonstrating creativity in its approach and methodology.

Dear reviewer,

thank you for your comments and suggestions. We tried to implement them into the text to improve the quality of out paper. We appreciate the time you took to review the text in such detail to help us make it better. Hereinbelow we address your points:

Suggested improvements:

Comment: 1. Provide a more explicit statement of how the findings might translate to real-world orthodontic practice.

Response: 1. This point was added into Discussion.

Comment: 2. Further analysis into the nature of specific facial changes, which caused an increase in the attractiveness scores.

Response: 2. We cannot determine a causal relationship. The facial changes were quantitatively analysed using Gimp, as described in the Methods section.

Comment: 3. Discuss more critically how biases in AI-enhanced photographs would influence orthodontic planning and treatment.

Response: 3. The point of AI bias was added into Discussion.

Comment: 4. Provide additional references for claims about AI capabilities in the introduction.

Response: 4. Additional references were provided.

Comment: 5. More in-depth detail about the specific FaceApp algorithms/filters used to better reproduce them.

Response: 5. The underlying algorithms are FaceApp know-how and cannot be explained in detail. However, more information was added so that readers can reproduce the research.

Comment: 6. Clarify why you analyze in detail only the five pairs with the greatest attractiveness difference.

Response: 6. This point was addressed in the text. The part of the study which analysed anthropometric points and distances, aims to correlate given face changes with the biggest attractiveness score changes. If we analyzed changes in all pictures and put them together, it would be less obvious which changes might have pointed towards the biggest attractiveness shift. For our upcoming paper we are currently developing a computer programme for automated analysis of facial landmarks. If successful, we will surely run all faces through this programme, and perhaps we would get some novel findings.

Comment: 7: Give some measure of interrater reliability of the attractiveness ratings.

Response: 7. Interrater reliability measure was calculated and added to the main text.

Comment: 8. Move some information from the discussion (e.g., about FaceApp's algorithms) to the methods section.

Response: 9. This suggestion was accepted and the Methods section was extended.

Comment: 9. Ensure the consistent use of first-person pronouns throughout the paper.

Comment: 9. We revised the pronouns.

Comment: 10. More citations should be added to some statements found in the introduction and discussion sections.

Response: 10. More references were added.

In conclusion, while the paper is highly promising and treats a critical topic, it needs moderate revisions in order to meet publication standards. Focusing on enhancements to the discussion section, clinical implications, and potential biases, organization of results, and improved writing, the authors could realize major improvements in the impact and clarity of their work. These improvements are what will make this innovative research find its value addition to the current knowledge on AI-assisted orthodontics.

We believe that we managed to incorporate all of your comments into this paper. Thank you for helping us improve the quality of this article.

Kind regards.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Revisions were not performed.

The title was not appropriate, but the authors refused to change it. The research does not focus on orthodontic treatment planning. The fact that the results of the present study could potentially be used for orthodontic treatment planning is, based on the present paper, only a possibility. I think it is unfair to the readers to have a title that is not appropriate for the manuscript and therefore it should not be accepted.

lines 47-50 "an orthodontist needs to remeber..." is appropriate for a book chapter where authors can provide their own opinion, but not for a scientific paper where authors should only provide scientific background.

1800 responders were invited and only 159 replied. this should have been mentioned in the article and discusses.

authors responded to my comment that the statistician defined 100 faces as needed for sample size. Despite the fact that this was not described in the manuscript, the present sample is half of what is needed; therefore the results do not have a statistically significance.

I feel the present article is of low scientific significance and should therefore not be accepted.

Author Response

Reviewer 1

Revisions were not performed.

The title was not appropriate, but the authors refused to change it. The research does not focus on orthodontic treatment planning. The fact that the results of the present study could potentially be used for orthodontic treatment planning is, based on the present paper, only a possibility. I think it is unfair to the readers to have a title that is not appropriate for the manuscript and therefore it should not be accepted.

lines 47-50 "an orthodontist needs to remeber..." is appropriate for a book chapter where authors can provide their own opinion, but not for a scientific paper where authors should only provide scientific background.

1800 responders were invited and only 159 replied. this should have been mentioned in the article and discusses.

authors responded to my comment that the statistician defined 100 faces as needed for sample size. Despite the fact that this was not described in the manuscript, the present sample is half of what is needed; therefore the results do not have a statistically significance.

I feel the present article is of low scientific significance and should therefore not be accepted.

 

Dear Reviewer,

Thank you for your further comments on our manuscript. We appreciate your continued engagement with our work. We would like to address the points you raised and also clarify the reasoning behind our decisions, while ensuring that our responses adhere to the ethical principles of scientific review and publication, as highlighted by the guidelines provided by MDPI.

Title of the Article: We understand that you have strong reservations about the title of our manuscript. As previously mentioned, our intention with the title was to reflect the potential application of AI-modified facial images in orthodontic treatment planning. However, we acknowledge that this might be interpreted differently by readers, and we appreciate your perspective on this matter. In the first round of the review, and after careful consideration, we have chosen to retain the title because we believe it encapsulated the clinical implications of our findings, which was the direction we intended to emphasize. We understand that there can be differing opinions on the appropriateness of a title, but we have made this decision based on the alignment with our research objectives and audience. We believe we have properly explained it in the first round and evaluated your comment as recommendation, not request we are obliged to follow.

Seeing your resilient point regarding the title of the article, we are changing the title to:

“The Potential of AI-Powered Face Enhancement Technologies in Orthodontic Diagnosis and Face-Driven Treatment Planning.“

We are sorry that you feel that “Revisions were not performed” and that you feel your previous points were not followed. We have addressed all the revisions you have suggested. Not implementing some with elaborated explanation. We have provided the original document with tracked changes and also other reviewers have evaluated our revisions as extensive. Also in this second round we address all your comments.

 

Your previous comment: Lines 38-41 are not scientific or appropriate

The more colloquial sentence “The cult of beauty does not readily forgive personal faux-pas, and if one wants to keep their clients, they must provide them with the most up-to-date services – when it comes to quality, ability to last, as well as aesthetics” was now rephrased to: “ In clinical practice, if one wants to meet patient expectations, they must provide them with up-to-date services – when it comes to quality such as dental and facial aesthetics and oral function.” We also added more sources to significantly strengthen the scientific underpinning of this part.

 

Your previous comment: Lines 47-50 are not needed

Lines 47-50: We respect your viewpoint regarding the appropriateness of the phrase "an orthodontist needs to remember" within the manuscript. Our intention was to underline the practical considerations that should be borne in mind by clinicians when interpreting the aesthetic outcomes of orthodontic treatment. We understand that such statements might be more common in a book chapter or a review article. Nonetheless, we have provided a scientific basis for this statement by referencing relevant literature. If deemed necessary, we are willing to rephrase this sentence to make it more suitable for a scientific manuscript while still conveying the intended message.

Unfortunately, we did not perceive this comment as your insisting for these lines to be removed. Nor did we understand until this second round based on that comment that you were not content with the phrasing “an orthodontist needs to remember” that you mention in this round. So thank you for the clarifications. This sentence is now rephrased. The more specific the comment, the more it helps us address your concern and implement your suggestions to improve the article.
To explain:
We are aware of diversity and the subjective aspect of facial beauty. Being professionals in the area, we need to appreciate both facial beauty and the occlusal stability concept – stable occlusion ensures stable results and thus needs to be respected. Therefore, we did our best to explain in the previous round: “We need to appreciate the need for both stable occlusion and smile aesthetics and facial harmony.” We hope that you understand why this idea is kept in the paper.

 

Your previous comment: How many people were invited to revise the pictures?

Response Rate (1800 Invitations, 159 Responses): We agree that the response rate is an important aspect of our study, and we apologize for not including this information in the manuscript initially. We will update the manuscript to include the details about the number of invitations sent and the response rate, along with a discussion on the potential implications of this response rate on the study's findings. We will also address how this response rate might affect the generalizability of our results and discuss potential biases.

You have evaluated statistical methods previously in the first round as “adequately described” - as you have ticked the box “Yes” for the question “Are the methods adequately described?“. Therefore, we only explained the number of respondents to you and did not add it to the article. However, as we appreciate your comment, the missing information is added to the article and discussed upon in this round to reach the aims you set.

Your comment: authors responded to my comment that the statistician defined 100 faces as needed for sample size. Despite the fact that this was not described in the manuscript, the present sample is half of what is needed; therefore, the results do not have a statistically significance

We understand your concern about the sample size being insufficient to achieve statistical significance. We have revisited the statistical analysis section to ensure that the sample size and its implications are clearly described in the manuscript. Additionally, we have elaborated on the limitations of our study, including the potential impact of the sample size on the robustness of our conclusions. We do rely on the well-regarded expertise and reasoning of our colleague, a co-author of this paper - Dominika Sónak Ballová from Department of Mathematics and Descriptive Geometry, who is a professional with extensive experience in statistics, both teaching at a university and performing statistical analyses for a living. Based on your previous comment, in our response we added “and based on chosen statistical methods, the number of generated faces was set to 100 pictures” to the article in the previous round. Hereinbelow I provide our statistician’s reply to your concern:
“The number of 50 faces is due to the used statistical methods and the number of respondents of the study is sufficient. The sample of 100 faces was meant as a paired sample, or in other words as a dependent sample (before and after the modification).”

We recognize the importance of maintaining high ethical standards in scientific research and publication, as outlined in the MDPI guidelines. We have made every effort to address your concerns with transparency and respect. We believe that scientific discourse should remain constructive, and we hope that our responses demonstrate our commitment to the integrity of our research and the publication process.

All in all, we perceive that our pilot study has a clear message that is supported by numbers and statistics. We aspire to study AI-powered face-enhancement technologies and their use in both orthodontic diagnosis and treatment planning further, however, hoping to inspire other research teams around the world. In the meantime, we believe this paper has a clear added value and a message to be shared with the community.

In conclusion, we believe that the research presented in this manuscript has value for the scientific community, particularly in its exploration of the intersection between AI technologies and orthodontic treatment. However, we also acknowledge the importance of presenting our findings with clarity and precision, and we are committed to making any necessary revisions to enhance the manuscript's quality.

Thank you once again for your time and effort in reviewing our work. We hope that our responses address your concerns and that our manuscript will be considered for publication.

Kind regards,

Authors

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Overall the authors seem to have made an effort to address all the raised concerns and, more importantly, implement most of the suggestions from the previous review. There are really serious improvements in the paper, especially in terms of clarification of methodology, depth in the discussion, and general organization. Few more:

1. Elaborate more on the specific facial changes correlated with attractiveness increases; maybe use visual supports or diagrams.
2. Include a discussion regarding possible limitations when using AI-generated faces in place of real patient photographs.
3. Elaborate on the ethics of deploying AI for use in treatment planning of orthodontics, particularly around patient expectations and consent.
4. Expand on how these results compare with standard measurements for orthodontic aesthetic assessment.
5. Maybe mention future research directions based on your findings briefly.

Author Response

Reviewer 3

Overall the authors seem to have made an effort to address all the raised concerns and, more importantly, implement most of the suggestions from the previous review. There are really serious improvements in the paper, especially in terms of clarification of methodology, depth in the discussion, and general organization.

 

Dear Reviewer,

Thank you very much for your kind reply. We are thankful you appreciate the efforts we put in implementing your suggestions and happy you deem our efforts successful. We feel your previous recommendations really helped to elevate the quality of the manuscript and glad to read you feel the same way. The suggestions you made are most welcome, and we hope we have managed to address them all also in this round.

Please allow us to elaborate how we improved the article as per your comments:

1. Elaborate more on the specific facial changes correlated with attractiveness increases; maybe use visual supports or diagrams.

We have added a chart showing the most prominent changes overall and discussed on the differences between the enhancement of male and female faces. We hope this addition is to your satisfaction.

2. Include a discussion regarding possible limitations when using AI-generated faces in place of real patient photographs.

We added a section discussing the disadvantages and limitations of using AI-generated faces in our study. In addition, we also explained why such methodology was chosen.

3. Elaborate on the ethics of deploying AI for use in treatment planning of orthodontics, particularly around patient expectations and consent.

Ethics, consent and patient expectations are areas of major concern for us too. Thank you for letting us elaborate in these areas. We added elaborations on privacy, data safety and expectations.

4. Expand on how these results compare with standard measurements for orthodontic aesthetic assessment.

We appreciate you highlighting this interesting direction of thoughts. We added a separate paragraph on the traditional and novel measurements.

5. Maybe mention future research directions based on your findings briefly.

We made sure to expand this area as per your suggestions.

In summary we really appreciate your praise and your valuable suggestions. We feel that we successfully implemented your recommendations and they added significantly to the value of the manuscript. All of these together made it possible to elevate the article to the very high level we always aspired it to be.

Kind regards,

The Authors

Author Response File: Author Response.pdf