Face Recognition Using Popular Deep Net Architectures: A Brief Comparative Study

Round 1

Reviewer 1 Report

Dear corresponding author,
the paper is well written but needs major revision before publication.
1) recent literature is full of similar research studies, as the ones I report below. Please highlight your contribution clearly in the text.

a - Bashbaghi, S., Granger, E., Sabourin, R., & Parchami, M. (2019). Deep learning architectures for face recognition in video surveillance. In Deep Learning in Object Detection and Recognition (pp. 133-154). Springer, Singapore.
b - Chaudhuri, A. (2020). Deep Learning Models for Face Recognition: A Comparative Analysis. In Deep Biometrics (pp. 99-140). Springer, Cham.
c - Shepley, A. J. (2019). Deep learning for face recognition: a critical analysis. arXiv preprint arXiv:1907.12739.

2) please discuss how and why feed-forward classification ANNs could not be used for face recognition. Consider in the reference papers like:

a - Bonfitto, A., Tonoli, A., Feraco, S., Zenerino, E. C., & Galluzzi, R. (2019). Pattern recognition neural classifier for fall detection in rock climbing. Proceedings of the Institution of Mechanical Engineers, Part P: Journal of Sports Engineering and Technology, 233(4), 478-488.
b - Singh, A. K., Tiwari, S., & Shukla, V. P. (2012). Wavelet based multi class image classification using neural network. International Journal of Computer Applications, 37(4), 21-25.

3 - Figure 3 must be revised since it has poor quality. The same applies to Figures 5 and 7.

4 - it would be very interesting to see examples of correctly classified images versus wrong classified images, with respect to different CNN architectures, in the Results section.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

The topic addressed by the authors sounds very interesting; authors gave an exhaustive general description of the networks involved in the study. I would suggest some improvements:

• Reversing paragraph 57-61 with paragraph 62-66 could make the text more fluent;
• Lines 160-161: authors could clarify if parameters are intended for object detection;
• Lines 364-366: could the authors give possible explanation?
• Authors tested the net architectures with predefined parameters; why didn’t they analyze the possibility of using pre-trained networks, perhaps for face recognition purposes?

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Dear corresponding author,

please make this changes prior to final submission.
1) Figure 3 must be revised and originally created.
2) the results shown in Figure 11 must be clearly commented and related conclusion must be clearly written in the main text.

Author Response

Please see the attachment.

Author Response File: Author Response.docx