An Analysis of Cholesteric Spherical Reflector Identifiers for Object Authenticity Verification

Round 1

Reviewer 1 Report

Arenas and colleagues prepared the cholesteric spherical reflector (CSR) arrays and tested them for use in the application of the object authenticity verification. CSR IDs’ images prepared by various methods were characterized using microscopic and macroscopic methods.

1. Many paragraphs are not clear and there are some typos. English correction is recommended before publications.
2. Many words are in italics. Authors should check the style of the MDPI journal.
3. The conclusion is not clear.
4. In the caption of Figure 1, what is the “optimal microscope”? Do the authors mean “optical microscope”?
5. Figures 3 and 4 are described earlier than Figure 2 in the text. Figure 2 should be explained first.
6. More explanation should be added to the caption of Figure 3.
7. On lines 54-55, in the sentence “The interactions… is…”, “is” should be “are”.
8. On line 70, delete “s” in “ands”.

Author Response

Dear reviewer, 

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper should be revised. There are many technical details and description of the verification procedures. However the introduction does not contain any physical description of the method. It should be also useful to provide the comparison of the proposed method with other methods of the verification, including the reliability of the results and time stability of the method. 

Author Response

Dear reviewer, 

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This article is about physical unclonable functions based on spherical cholesteric shells. In comparison to their previous works, the authors are now discussing larger and more realistic datasets, using both simulated noisy replica and real-life noisy replica. Most importantly, they also show the feasibility of using relatively low-cost equipments for challenge verification.  I have a few minor comments, which are detailed below.

The authors are not really clear about the origin of noise for the Dino-Lite acquisitions: how can they ensure that different out-of-focus images are taken for the same replica set, is there a translation stage that can move the sample and/or camera? I also wonder whether deblurring algorithms could be added to the alignment step, in order to improve the robustness with respect to out-of-focus effects.

It is not really clear what the authors mean by "The illumination was perpendicular to the sample", is there a reflection cube in the Dino-Lite microscope (like in a real reflection microscope)? Probably it would not be too far-fetch to imagine a real-life validation scenario would involve a controlled illumination spectra (potentially with additional "noisy" environment light sources), here I am not sure what the illumination conditions actually are and why the contrast is so low (CSRs are excellent reflectors since it relies on a photonic bandgap...). It should certainly not depend on the black background since light is reflected by the CSRs themselves. Maybe the illumination source is not powerful enough at the relevant wavelengths? Could the authors provide more details on this point?

Author Response

Dear reviewer, 

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors still did not follow my comments properly. The physics of the phenomena is not clear as well as the comparison of the application of this method with other methods for object authenticity verification is not provided.

Author Response

Dear reviewer, 

We apologize that our answer did not correctly address your comments. 

We have modified Sections 1 and 2 accordingly. However, we would like to clarify the two comments you pointed out:

• The physics of the phenomena is not clear.

In this comment, we did not fully understand to which of the following points you refer to: 

1. how we envision the verification to happen when we handle an object carrying a CSR ID, OR
2. the physical methods of production of CSRs?"

For (1), we can say that we have not defined yet an authentication mechanism with the different entities. But we built a dataset that roughly simulates the real conditions. In general, the read-out of the images was done in a “controlled” environment, i.e., a laboratory environment condition.

If you refer to (2), we added lines 53-65 for describing the physical methods of the CSRs production. Thus, we summarized the production process of CSRs; we also preferred not to give too many details about the production phase as the current scope of the paper is different.

• The comparison of the application of this method with other methods for object authenticity verification is not provided.

We completely agree that a comparison between the different technologies must be included. We renamed Section 2 (CSRs in object verification and comparison with other authentication technologies) and we added a comparison with similar authentication technologies. The changes can be tracked from lines 129 to 183.

We would like to acknowledge your pertinent comments for improving the paper.

 

Yours faithfully

Mónica Arenas

Author Response File: Author Response.pdf