Next Article in Journal
Can We Use Satellite-Based FAPAR to Detect Drought?
Next Article in Special Issue
Intelligent Framework for Learning Physics with Aikido (Martial Art) and Registered Sensors
Previous Article in Journal
Nonlinearity Correction in OFDR System Using a Zero-Crossing Detection-Based Clock and Self-Reference
Previous Article in Special Issue
Physical and Tactical Demands of the Goalkeeper in Football in Different Small-Sided Games
Article
Peer-Review Record

Dancing Salsa with Machines—Filling the Gap of Dancing Learning Solutions

Sensors 2019, 19(17), 3661; https://doi.org/10.3390/s19173661
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Sensors 2019, 19(17), 3661; https://doi.org/10.3390/s19173661
Received: 19 June 2019 / Revised: 6 August 2019 / Accepted: 21 August 2019 / Published: 23 August 2019
(This article belongs to the Special Issue Advanced Sensors Technology in Education)

Round 1

Reviewer 1 Report

In this paper, the authors must differentiate: (a) the novelty of their approach or (b) the system’s innovation. In the Windows Store are several applications (e.g., Dance Central) similar to the one presented in this paper.

The authors mention that the Kinect sensor is used to “track the learner’s facial expression and body” – but, here is the only place where the facial expression is mentioned. Please present more data if the system uses facial expression.

Please use a higher font in Figure 3 – the text, on the printed paper, is difficult to read.

Please, also, detail the technical aspects un “Salsa steps recognition” subchapter.

Author Response

Dear reviewer,


thanks for taking the time to review our manuscript and providing us for valuable feedback to improve it. 



Point 1: “In this paper, the authors must differentiate: (a) the novelty of their approach or (b) the system’s innovation. In the Windows Store are several applications (e.g., Dance Central) similar to the one presented in this paper.

Response 1: Thank you for the comment, in the revised version of the manuscript we addressed this point by refining our introduction (Lines 18-35) explaining how our system is different from available technologies that can help people to learn how to dance. Furthermore, we also modified a bit the conclusion to make this point more explicit (Lines 394-396 ).



Point 2: “The authors mention that the Kinect sensor is used to “track the learner’s facial expression and body” – but, here is the only place where the facial expression is mentioned. Please present more data if the system uses facial expression.

Please, also, detail the technical aspects un “Salsa steps recognition” subchapter.
Response 2: Thanks for the comment, indeed we agree that a better explanation of how the system recognizes different aspects of the learner was required to improve the paper. To address this, we added the subchapter 2.4 (lines 178-203) where we explain how all salsa dancing mistakes are analyzed including the facial recognition (Line 186) with the kinect sensor. 

We also added chapter 3: (Lines 204-256) where we reported the technical evaluation of the used recognition techniques including deeper details into the salsa steps recognition 



Point 3: 

Please use a higher font in Figure 3 – the text, on the printed paper, is difficult to read.

Response 3: We increased the font of Figure 3, thanks for the suggestion.





Reviewer 2 Report

The paper presents a first version of a system aimed at supporting the learning and practice of dancing, called Dancing Couch (DC). The current version deals only with eight basic steps of Salsa. As many other dancing tutoring system, the system described in the paper uses a depth camera to capture the movement of the dance trainee. For recognizing the Salsa steps, the authors used Kinect Studio and Kinect Visual Gesture Builder.

 

The paper also presents a subjective user evaluation of the system.

 

The authors argue that the system allows the addition of other dancing styles.

 

In my opinion, the DC system is in an initial phase of development, precluding the publication of the paper.  It is not possible to identify any relevant contribution and novelty in the paper. As point out in manuscript, there exist systems that seek the same goal and use the same sensor (depth camera). However, the paper does not compare the DC with these other solutions, and therefore fails to present convincing arguments for its novelty and improvement over the other approaches.

 

The range of the gestures handled by the DC system is too limited – only basic 8 steps of on dance style. The argument that the system allows the addition of other dancing styles is not supported by concrete example. A discussion on the procedure and difficulties to extend the system to cope with other dancing style would significantly improve the quality of the paper.

 

The subjective/qualitative evaluation presented in the paper has its merit, but together with an objective/quantitative evaluation would permit a better assessment of the system. For further development and evaluation of the system, I would suggest the use of a control group, whose components do not use the DC system, besides the experimental group.  The use of a control group would allow the objective measurement of the effect of the usage of the DC system.


Author Response

Dear reviewer,


thanks for taking the time to review our manuscript and providing us for valuable feedback to improve it.



Point 1:In my opinion, the DC system is in an initial phase of development, precluding the publication of the paper.  It is not possible to identify any relevant contribution and novelty in the paper. As point out in manuscript, there exist systems that seek the same goal and use the same sensor (depth camera). However, the paper does not compare the DC with these other solutions, and therefore fails to present convincing arguments for its novelty and improvement over the other approaches.

Response 1: Thanks for the feedback, in the revised version of the manuscript we addressed this point by refining our introduction (Lines 18-35) explaining how our system is different from available technologies that can help people to learn how to dance, and made modifications in Lines 53-58 to explicitly point out the contribution of the paper. Furthermore, we also modified a bit the conclusion to make this point more explicit (Lines 394-396 ).



Point 2:The range of the gestures handled by the DC system is too limited – only basic 8 steps of on dance style. The argument that the system allows the addition of other dancing styles is not supported by concrete example. A discussion on the procedure and difficulties to extend the system to cope with other dancing style would significantly improve the quality of the paper.

Response 2: Thanks for the suggestion. To address this we added the evaluation of basic side steps (228-237), showing how the system can be expanded to recognize more steps. Further, we added subsection 2.1 (Lines 108-125)  about the architecture and how it can be used to expand the DC. In the discussion section (Lines 374-384), we discuss the procedure to expand the DC to different dancing styles.



Point 3:The subjective/qualitative evaluation presented in the paper has its merit, but together with an objective/quantitative evaluation would permit a better assessment of the system. For further development and evaluation of the system, I would suggest the use of a control group, whose components do not use the DC system, besides the experimental group. The use of a control group would allow the objective measurement of the effect of the usage of the DC system.

Response 3: As this is the first version of the DC, for this study we followed the research practice suggested in (Nieveen & Folmer,2013) to first conducting a formative evaluation of the system, identify the how users experience it and identify challenges for using the system. We consider that this formative evaluation provided us with enough information to improve the system. Once the system is improved indeed to confirm its efficacy a summative assessment of it with a control group would be suitable.



Nieveen, N., & Folmer, E. (2013). Formative evaluation in educational design research. Design Research,153, 152-169.



Reviewer 3 Report

The paper, as stated by the authors in the abstract, would present the first implementation od a generic system to support the practice of Salsa dancing and to provide a user evaluation of the Dancing Coach (DC).

in the way the paper has been structured, it appears that only the last issue has been addressed as clearly indicated in the results section . In fact the results show the evaluation given by the  pre- and post- questionnaires while some aspects of the system are described in general manner in the paper.

If the intenction of the authors was to present the entire system, the methods to extract movement and facial expression during the coaching phase and a more accurate description should be provided in the methods and in the related results.

Several aspects related to motion recognition, feature extraction and classification should be given.
The inaccuracies related to the foot tapping phases (right and left) should be discussed.

It is not clear how the system can recognise the facial expression (as mentioned in Dancing coach application).

Author Response

Dear reviewer,

thanks for taking the time to review our manuscript and providing us for valuable feedback to improve it. 



Point 1:The paper, as stated by the authors in the abstract, would present the first implementation od a generic system to support the practice of Salsa dancing and to provide a user evaluation of the Dancing Coach (DC).



in the way the paper has been structured, it appears that only the last issue has been addressed as clearly indicated in the results section . In fact the results show the evaluation given by the  pre- and post- questionnaires while some aspects of the system are described in general manner in the paper.



Response 1: Thanks for the comment. We addressed this comment by refining our introduction showing how the DC is different from current existing solutions (Lines 18-35). We also added subsection 2.1 (Lines 108-125)  about the architecture and how it can be used to expand the DC. In the discussion section (Lines 374-384), we discuss the procedure to expand the DC to different dancing styles.



Point 2:If the intenction of the authors was to present the entire system, the methods to extract movement and facial expression during the coaching phase and a more accurate description should be provided in the methods and in the related results.



Several aspects related to motion recognition, feature extraction and classification should be given. 

The inaccuracies related to the foot tapping phases (right and left) should be discussed.



It is not clear how the system can recognise the facial expression (as mentioned in Dancing coach application).



Response 2: Thanks for the comment, indeed we agree that a better explanation of how the system recognizes different aspects of the learner was required to improve the paper. To address this, we added the subchapter 2.4 (lines 178-203) where we explain how all salsa dancing mistakes are analyzed including the facial recognition (Line 186) with the kinect sensor. 

We also added chapter 3: (Lines 204-256) where we reported the technical evaluation of the used recognition techniques including deeper details into the salsa steps recognition



Round 2

Reviewer 1 Report

In this form, the work is much better than the previous one.

Although the response received refers to another program than that the one pointed by me, I believe that the authors responded, at least partially, to my question.


Author Response

Dear Reviewer,

 

thanks again for taking the time to review our manuscript. we are happy that we could address your feedback to improve the paper.

Reviewer 2 Report

I am still of the opinion that the DC system is in an initial phase of development, precluding the publication of the paper.

The main drawbacks of the manuscript are: 1) the range of the dance steps handled by the DC system is too limited; 2)  the approach for recognizing dance steps uses the Kinect Studio and Kinect Visual Gesture Builder and does not seem to be novel; 3) the manuscript does not provide convincing evidence that the system can be easily extended to incorporate new dance styles; 4) the recognition strategy does not seems robust enough to cope with more complex dance steps that include occlusions and twists, for example; 5) the smile recognition approach also uses built-in functions of the Kinect SDK and therefore is not novel; 6) the user evaluation does not provide empirical evidence of the effectiveness of the system for the problem at hand.

My suggestion to the authors is to continue the development and in the future submit an improved version of the manuscript with more robust and convincing results

Author Response

Dear reviewer, 

thanks for taking the time to review our manuscript and providing us for valuable feedback to improve it.

Point 1: the range of the dance steps handled by the DC system is too limited

Response 1: We are aware of that and mentioned it in line 371-373. However, basic dance steps are crucial to learn e.g. for more complex movements like twists (line 55-57).

 

Point 2: the approach for recognizing dance steps uses the Kinect Studio and Kinect Visual Gesture Builder and does not seem to be novel

Response 2: To the best of our knowledge there is no other system that uses Kinect Studio and Visual Gesture Builder to recognize dance steps with the Kinect v2 as presented in the manuscript. That is why we introduced other systems in line 35-46 that mostly used the older Kinect and therefore not Visual Gesture Builder.

 

Point 3: the manuscript does not provide convincing evidence that the system can be easily extended to incorporate new dance styles

Answer 3: As we mentioned, we think that the architecture can facilitate expanding the DC in line 108-126 e.g. by following the same procedure as presented in the manuscript (line 374-376) and discussed possible obstacles in line 377-382.

 

Point 4: the recognition strategy does not seems robust enough to cope with more complex dance steps that include occlusions and twists, for example

Response 4: We are aware of that and mentioned it in line 351-358. However, the recognition was built for basic dance steps that are crucial to learn e.g. for more complex movements like twists (line 55-57).

 

Point 5: the smile recognition approach also uses built-in functions of the Kinect SDK and therefore is not novel

Respone 5: That is true but we would like to point out that the argument of the manuscript is not about a novel approach for smile recognition with the Kinect v2.

 

Point 6: the user evaluation does not provide empirical evidence of the effectiveness of the system for the problem at hand

Response 6: Our overarching research on the DC follows a design-research based methodology (Anderson & Shattuck, 2012) which is an iterative process. In the manuscript, we present the first full iteration on the DC. As you point out, indeed this means that the DC is currently at an early developmental stage. However, we consider that the presented study provides already valuable information on the general user experience and potential of the DC, and crucial information on how to improve it and scale it. Thus, in our view, the manuscript already presents a substantial enough contribution to be published.

We agree with you in your concerns that the DC has a lot of room for improvement and that later iterations of our overarching research of the DC will allow us to provide manuscripts with more substantial contributions. To address this we would first need to improve the DC based on the results obtained in this study, come up with a new study design to test the effectiveness of the DC, and conduct a new set of user tests. All of this consists of a completely new study, which is outside of the scope of the presented manuscript and it is not feasible to include this in short time. 

Anderson, T., & Shattuck, J. (2012). Design-based research: A decade of progress in education research?. Educational researcher, 41(1), 16-25.

Reviewer 3 Report

Authors followed my suggestion. The paper can be considered for pubblication.

Author Response

Dear Reviewer,

 

thanks again for taking the time to review our manuscript. we are happy that we could address your feedback to improve the paper.

Back to TopTop