Electromyogram (EMG) Signal Classification Based on Light-Weight Neural Network with FPGAs for Wearable Application

Round 1

Reviewer 1 Report

In the paper entitled "EMG Signal Classification Based on Light-Weight Neural Network with FPGAs for Wearable Application", the author proposes to use a neural network with properties that allow its implementation in the edge device.

The author assessed the performance of the proposal by applying it to a dataset of EMG. Six hand movements were classified with high values of accuracy and area under the ROC curve. 

The manuscript is well-written and presents a vast explanation and analysis of the different variables and parameters involved in the study.

I think the work is of interest in the area of health care since the proposed methodology can be used for diagnosis without much computational effort.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear author

In the present research titled: “EMG Signal Classification Based on Light-Weight Neural Net-2 work with FPGAs for Wearable Application”, firstly, I want to congratulate you for the very well written, and organized paper. Additionally, I still have some suggestions related to the manuscript's aesthetics and theoretical points.

 In this sense, I have the following suggestions:

Title

Please, try not to use the abbreviation of terms in the title, because it can confound readers just at the beginning of the text

Keywords 

Avoid repeating the same terms of the title in the abstract keywords, it should look redundant and not help readers to find your article in the web databases

Introduction

There is a lack of citations in the introduction, please, look for citations that contemplate the information stated in the paper introduction.

Figure 1

Apparently, figure 1-a has the x-axis label font bigger than in x-axis of the figure 1-b, rearrange the figure, please. Therefore, it's possible to improve the quality of all figures, so, look for ways to make this.

Conclusions and discussions

I perceive that you should enjoy better the part where you say about the benefits of the device for the practice area, however, you should give more practice examples (lines 389-396).

No further comments.

Best regards; 

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this paper, author presents the classification of EMG signal based on Light-Weight Neural Network classifier. The signal processing and the classification algorithm are implemented using FPGA technology. 

This is a interesting work, however, I have some comments and remarks.

- Abstract needs to be rewritten to understand the summary of the work done highlighting the principal contributions.

- In the introduction, the comparison with previous works must be more precise in order to highlight the real contribution of this work. In addition, the motivation and background of wide practical use of the theoretic results presented should be clearly emphasized to facilitate the readers.

- Some figures are not clear. Please provide a high-resolution figure and check the selected colors.

- English is generally good, but needs to be polished further. The manuscript should be formatted better and some spelling and grammar should be checked carefully.

- The selection of the related useful feature extraction approaches should be more discussed. How the authors chose these approaches only based on the literature? I think that it's better to test some features and select the more suitable approaches. 

- Muscle selection and electrode placement should be more detailed and justified. 

- The selection of the sampling frequency should be justified (Shannon theory). I think that 500hz is too low for accurate signal reconstruction. 

-What kind of filter has been used (IIR?) for EMG preprocessing ?

- Experimental platform is missed. EMG sensor, electrode types, acquisition system and other components should be presented and discussed. 

- Discussion part is missed, it should be added. 

- Conclusion does not reflect the presented work. It should be rewritten by adding the limitations and perspectives.

Concluding, the whole paper is a combination of existing technologies without any real contribution, any real photo showing the experiment ! 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

1. Please give the details of proposed method for proposed model. I suggest the authors amend the calculation of your size of proposed method and the details is important for proposed method. 
2. The section 1 and section 2 should be merged as introduction. Are Figure 1-Figure 3 from existing literature? If yes, please give the reference situations. I suggest no figures in introduction section. 
3. In Figure 4, I suggest use figure for depict the details and information by Visio software. The data in Figure 5 is too simple, and I suggest the authors can amend related data in Figure 5 for comparison. 
4. In section 3 and section 4, the content of experiments needs to amend related experiments to compare related SOTA in recent three years. I recommend the authors amend related experimental results of proposed method of SOTA according to the published paper in IEEE, Springer and Elsevier. 
5. Please update references with recent paper in CVPR, ICCV, ECCV et al and Elsevier, Springer. In your section 1 and section 2, I suggest the authors amend several related literatures and corresponding references in recent years. For example: FFTI: Image Inpainting Algorithm via Features Fusion and Two-Steps Inpainting (Journal of Visual Communication and Image Representation, https://doi.org/10.1016/j.jvcir.2023.103776). 
6. In the conclusion section, the limitations of this study and suggested improvements of this work should be highlighted.
   

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

All my suggestions have been implemented in the paper, I have no further comments. 

Author Response

Thank you for your kind response, and I will try to submit better papers in the future.

Reviewer 4 Report

 According to the revised manuscript and response letter, I recommend the revised manuscript for the journal.

Author Response

Thank you for your kind response, and I will try to submit better papers in the future.