PriRepVGG: Privacy-Preserving 3-Party Inference Framework for Image-Based Defect Detection

Round 1

Reviewer 1 Report

Comments

The presented work is useful and reviews some important shortcomings that have been presented appropriately, however some notes that would like to be presented to expand the field of study and research material in this manuscript:

§  The frameworks to address the deficiencies in confidentiality and privacy aspects were not presented very well in the theoretical framework and approach used in this study. Especially by competitors and malicious adversaries. Please focus on this aspect because of its importance to increasing the chance of the approach acceptance and theoretical framework of this study.

§  A number of studies are somewhat outdated, some of them are more than 10 years old, although the theoretical framework for this study has been greatly enriched during the past five years. Please take this aspect into consideration when selecting the research references for this study.

Author Response

Thanks for your comments! Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

An experiment in medicine, as mentioned in the summary, would have been interesting

Author Response

point 1:An experiment in medicine, as mentioned in the summary, would have been interesting

Response 1:Thank you for your valuable suggestions. We will expand the application and practice of this direction in our subsequent research

Reviewer 3 Report

Authors have proposed PriRepVGG to secure inference of diamond substrate defects under data server, model server, and compute server conditions. The following improvements were required in the paper 

·        In related work, there is no information that can be used to evaluate the author's actual work. There is a lack of proper formatting in the references. To justify this research, the author should include a literature review section and compare the pros and cons of existing research. 

·        Poor formatting, including tables, text, and paragraph positioning, and grammatical errors such as the use of words such as (we, our)

·        There is a need for more evidence and results to support the findings. In this case, the number of results is not sufficient, which does not result in the desired outcome.

·        There is no connection between the abstract and conclusion statements; they are two separate statements. A mixed conclusion can be drawn regarding future work and the methodology section.

·        The results demonstrate time and communication rather than a model of privacy, so the title of the paper is inaccurate.

 

 

Author Response

Thanks for your review, please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

I appreciate your response to my comments. The responses have been satisfactory to me. 

It is recommended that the authors refer to the following papers for more recent work in the field of privacy. 

Bazai, S. U., Jang-Jaccard, J., & Alavizadeh, H. (2021). Scalable, high-performance, and generalized subtree data anonymization approach for Apache Spark. Electronics, 10(5), 589.

Bazai, S. U., Jang-Jaccard, J., & Alavizadeh, H. (2021). A novel hybrid approach for multi-dimensional data anonymization for apache spark. ACM Transactions on Privacy and Security, 25(1), 1-25.

Lim, J. S., Hong, M., Lam, W. S., Zhang, Z., Teo, Z. L., Liu, Y., ... & Ting, D. S. (2022). Novel technical and privacy-preserving technology for artificial intelligence in ophthalmology. Current Opinion in Ophthalmology, 33(3), 174-187.

Jiang, H., Li, J., Zhao, P., Zeng, F., Xiao, Z., & Iyengar, A. (2021). Location privacy-preserving mechanisms in location-based services: A comprehensive survey. ACM Computing Surveys (CSUR), 54(1), 1-36.