Siamese Neural Network for User Authentication in Field-Programmable Gate Arrays (FPGAs) for Wearable Applications

Round 1

Reviewer 1 Report

I think this paper has some value but some issues need to be solved:

1  please add more theoretical analysis of your proposal. Give the hints on your underlying idea.

2 no comparison has been given for the related work, please add this part.

3 some related work on traditional authentication can also be discussed to present your novelty 

4 the typical application setting can be added such as how to construct App for implementing your idea

5 some work on authentication can be selected referred such as

Vehicle and Pedestrian Detection Algorithm Based on Lightweight YOLOv3-Promote and Semi-Precision Acceleration. IEEE Trans. Intell. Transp. Syst. 23(10): 19760-19771 (2022)

Cryptanalysis of a public authentication protocol for outsourced databases with multi-user modification. Inf. Sci. 488: 13-18 (2019)

Some problems are exist in the paper 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

1) What is the primary motivation for using electromyogram (EMG) signals for user authentication, and how do they compare to traditional methods like passwords and fingerprints in terms of security and user convenience?

2) Can you provide more information about the Siamese network and how it is employed for user authentication in the proposed system? How does it address the challenges of limited learning data in the wearable environment?

3) What is the significance of implementing the Siamese network on field programmable gate arrays (FPGAs) for wearable user authentication? How does this affect the system's performance and efficiency?

4) Could you explain in detail how the maximal overlap discrete wavelet transform (MODWT) method is used for time series data analysis and how it contributes to improving the accuracy of EMG signal-based authentication?

5) What is the size and complexity of the Siamese network implemented in the FPGA-based edge devices, and how do these factors impact the system's performance and resource utilization?

6) You mentioned the possibility of improving accuracy using multimodal techniques or long short-term memory (LSTM). Could you provide more information on how these techniques might be integrated into the proposed system and their potential impact on accuracy and security?

7) What are the limitations of the proposed user authentication system based on EMG signals, and how do these limitations impact its real-world applicability and performance, especially in terms of scalability and robustness?

8) How does the proposed bit quantization method work, and what are its implications for resource usage and the overall system's power consumption and security?

9) What are the key contributions of this research, and how does it advance the field of user authentication, particularly in wearable applications? Are there any potential future research directions or practical implications that should be highlighted?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this paper, the author proposes a novel user authentication method based on biosignals, specifically electromyogram signals. An  overall hardware structure and verification process through high-level synthesis for hardware deployment. The results are presented by means of precise figures and tables.

 

My overall evaluation of the paper is positive. English is acceptable, the methodology is well detailed, and I appreciate the detail given in the description of the tools used. I am in favor of accepting this paper, but just formulate minor comments that don't deserve a new round: perhaps the contributions of the paper must be explained in a more concise way in the Abstract and Introduction.

Nice English. Perhaps, a small polishing is welcome, but the English remains at a professional level.

Author Response

 Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Many faults is the paper, abstract need to.be improved, introduction weak,short and unclear. Related work is missing.

No expermintal work to verify the work. The discussion was shortened with the conclusion. 

 

English editing is required.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The revisions have substantially improved the paper, addressing previous concerns and enhancing its overall quality. I recommend accepting it for publication.

Author Response

Dear Reviewer,

I greatly appreciate your valuable comments. I will strive to submit an even more advanced version of the paper in the future.

Sincerely yours,

Hyun-Sik Choi.

Reviewer 4 Report

Lack of significance , despite the authors effort, I still can see many faults in the paper structure. 

 

The previous comments were not addressed. 

Some sentances are weak and need to be edited 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 3

Reviewer 4 Report

The comments were not addressed, some minor changes has been made which is not enough. 

English editing required by professional english editor