Multi-Output Sequential Deep Learning Model for Athlete Force Prediction on a Treadmill Using 3D Markers

Round 1

Reviewer 1 Report

he manuscript looks at Multi-Output Sequential Deep Learning Model for Athletes' Force Prediction on a Treadmill using 3D markers.
Before the paper can be considered for publication, the following revisions are required.
1. The research gap needs to be explicitly highlighted in the literature review along with the aim and scope of the study
2. The conclusion has to be rewritten to highlight key findings and achievement of the study
3. What is the practical implication of this work. Please discuss this in the Section 4 of the manuscript
4. Will this proposed study be easily adopted by other researchers to enhance biomechanics performance? Please explain in detail in section 4

Author Response

The point-by-point response is presented in "Review RNN 1.pdf", where a table with columns: "Comment from Reviewer," "Location in the Document (Section, Page, Lines)," "Comments from Authors" explain in detail the modifications made to the manuscript according to the reviewer's comments.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript applsci-1685947 presented an RNN trained based on a biomechanical dataset of regular runners that measures both kinematics and kinetics, the model will allow analyzing, extracting, and drawing conclusions about continuous variable predictions through the body, which then are passed to marking around different anatomical and reflective points (96 in total, 32 per dimension) that will allow the prediction of forces (N) in the three-axis (Fx, Fy, Fz), extracted from a treadmill with a force plate, on different velocities (2.5 m/s, 3.5 m/s, 4.5 m/s). To obtain the best model, a grid search of different parameters was done, by combining different types of layers (Simple, GRU, LSTM), loss functions (MAE, MSE, MSLE), and sampling techniques (Down-Sampling, Up-Sampling), the best performing model (LSTM, MSE, Down-Sampling) achieved an average coefficient of determination of 0.68, although when excluding Fz it reached 0.92. My overall impression of this paper is that it is in general well-organized. It was a pleasure reviewing this work and I can recommend it for publication in Applied Sciences after a major revision. I respectfully refer the authors to my comments below.

1. The English needs to be revised throughout. The authors should pay attention to the spelling and grammar throughout this work. I would only respectfully recommend that the authors perform this revision or seek the help of someone who can aid the authors. For example,

---(Line 5) The term “RNN” should be given its full name before its abbreviation.

 

2. (Page 2, last paragraph in the Introduction part) The “main contributions” is best to list clearly by breaking it down into three points. The reader can understand your contribution easily.
3. (Page 1, Section 1-Introduction) The original statement “… when compared to its use in the fields of robotics …” is suggested to revised as “… when compared to its use in the fields of robotics (1) MFDNet: Collaborative Poses Perception and Matrix Fisher Distribution for Head Pose Estimation (2) ARHPE: Asymmetric Relation-aware Representation Learning for Head Pose Estimation in Industrial Human-machine Interaction, …”.
4. In Section 2.2 (Loss functions), the reviewer suggests authors add some formal descriptions of the proposed model, such as loss functions, so that the reader can better understand the process.
5. (Page 10, Figure 5) The reviewer suggests authors introduce clearly the overall flow of Figure 5 (in the body or in the picture description).
6. Experimental pictures or tables should be described and the results should be analyzed in the picture description so that readers can clearly know the meaning without looking at the body. For example, describe the colorful markers in Figure 6 and describe the results of the analysis of this phenomenon.
7. (Page 2, Section I Introduction) Please add some related references. The original statement “… trying other ML algorithms or DL architectures, as Convolutional Neural Network (CNN), usually used to analyze images, could also be used in time-series analysis [17].” is suggested as “… trying other ML algorithms (DOI: DOI: 10.1109/TNNLS.2021.3055147) or DL architectures (EDMF: Efficient Deep Matrix Factorization with Review Feature Learning for Industrial Recommender System), as Convolutional Neural Network (CNN) (Anisotropic angle distribution learning for head pose estimation and attention understanding in human-computer interaction), usually used to analyze images, could also be used in time-series analysis [17].”
8. (Tables 3-4) Add a new table to demonstrate the scores of all the comparing methods. And the best scores should be bolded.
9. (Page 1, Introduction) The reviewer suggests authors don't list a lot of related tasks directly. It is better to select some representative and related literature or models to introduce with certain logic. For example, the latter model is an improvement on one aspect of the former model.
10. (Line 81) “… various formulae such as Mean Absolute Error (MAE), Mean Squared Error (MSE), …” is suggested as “… various formulae such as Mean Absolute Error (MAE) [1], Mean Squared Error (MSE) [2], …”. ([1] CARM: Confidence-aware recommender model via review representation learning and historical rating behavior in the online platforms [2] NGDNet: Nonuniform Gaussian-label distribution learning for infrared head pose estimation and on-task behavior understanding in the classroom)
11. The authors are suggested to add some experiments with the methods proposed in other literatures, then compare these results with yours, rather than just comparing the methods proposed by yourself on different models.

My overall impression of this manuscript is that it is in general well-organized. The work seems interesting and the technical contributions are solid. I would like to check the revised manuscript again.

Author Response

The point-by-point response is presented in "Review RNN 2.pdf", where a table with columns: "Comment from Reviewer," "Location in the Document (Section, Page, Lines)," "Comments from Authors" explain in detail the modifications made to the manuscript according to the reviewer's comments made.

Author Response File: Author Response.pdf