A Novel Transmittance Vis–NIR Hyper-Spectral Imaging Scanner for Analysis of Photographic Negatives: A Potential Tool for Photography Conservation
, Andrea Casini 1, Lorenzo Stefani 1, Barbara Cattaneo 1,2 and Marcello Picollo 1
Round 1
Reviewer 1 Report
The authors have developed a prototype for scanning photographic negatives using spectral imaging techniques in the visible and near-infrared range (400-900 nm) focused on the conservation of this heritage. The proposal is very interesting, the introduction gives sufficient information on the state of the art. However, as the experimental part progresses and, above all, when reaching lines 261-266, the reader realises that the title was too promising for the TRL achieved. The description of the results is too introductory. In fact, there is no results data in the abstract. No spectra are shown to assess the profile of the different parts of the negatives. Since there is no reference data and only an unsupervised analysis (PCA) has been done, it would be interesting to at least have a visualisation of all the principal components (PC) of the same image, indicating, in each case, the percentage of variance explained by each principal component and, also, the loadings of these PCs. This is the only way to determine whether a spectral technique is useful or not, or whether a simple desktop scanner could achieve the same results.
Apart from this general impression, there are many other aspects that, in my opinion, should be reviewed:
- Line 12. A gap between 900 and nm is missing. This happens a lot throughout the article, where a space between magnitude and units is missing.
- Line 16. There is a missing full stop after sector.
- Line 24. This section is not correctly numbered.
- Line 137: I assume that CH refers to "Cultural Heritage". If so, this abbreviation should be well defined.
- Line 177: What is the light-diffuser white screen made of? Has the possible effect of the absorption of this device in the infrared region been studied? What was the distance between the lamp and this diffuser, and were there any temperature problems due to the presence of halogen lamp?
- Line 180: I think the conversion between millimetres and micrometres is not correct.
- Line 183: Wasn't a measurement of dark current noise made? It is very usual (practically essential) in HSI.
- Line 227: Figure 3 is unnecessary.
- Line 286: It would be interesting to know how the sRGB image has been obtained from the hypercube: Have three discrete wavelengths been selected or has each band been obtained by filtering the spectral information as would be the case with the development of the negative? This is very important, as this is the only way to justify the use of spectral imaging instead of simply obtaining the image with a desktop scanner. It would be very illustrative to explain this colorimetric treatment.
- Line 304: The authors talk about the relevance of PC4. Which wavelengths are responsible for this main component? Are the loadings more important in the visible or in the infrared?
- In Line 334, the authors state: "In Figure 6 the PC6 of the examined film strip is reported, with a magnification of the first photogram…". However, in the caption of figure 6 it is stated that what is expressed is PC4.
- Line 379: The link includes part of the subsequent text and does not work properly.
- I would change the title in order to better match the title with the results obtained.
For the reasons stated above, a "major revision" of this article is recommended. Although this revision may seem very critical, I would like to stress again how interesting the proposal is. I strongly encourage the authors to consider the points raised, as it could become a reference article in this scope.
Author Response
We gratefully acknowledge the Reviewer for the insightful comments and the fruitful suggestions. We have addressed most of the suggestions and in some cases we have explained our reasons for rebuttal.
The Reply point to point is reported in the attached file. Please see the attachment for the detailed list of changes.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This paper introduces a transmittance prototype HSI camera, operating in the 400-1000nm range; the system is designed to scan photographic negatives films and slides. The authors briefly describe the spectral camera and the detection approach; in addition, some test measurements and preliminary data analysis are provided. HSI datacubes were acquired on a set of degraded photographic negatives, and analyzed with multivariate algorithms, such as PCA, which enabled to obtain enhanced images and recover details apparently lost in the damaged original.
The paper is very interesting, how ever there are some aspects that should be addressed before publication:
1. The description of the setup lacks information about the objective (focal distance, aperture,...). In addition, very little is disclosed about the mechanical movement of the camera. For example, which motors were used? And how important is the distance from the sample? What is the field of depth? Did the authors also implement a system to keep the camera at a constant distance from the film?
Concerning the spatial sampling step, the authors claim that it is “0.04 mm (0.37 um)”. I think it should be instead 37 um.
2. The authors mention that the measurement mode, originally reflectance, was changed to transmittance: such changes involved the optical part (spectrometer, objectives, detectors,...) or simply the mechanical frame of the system? I do not expect relevant changes in the detection path switching from reflectance to transmittance, but if any, the authors should discuss the main aspects of such modifications. With the new system, what is the acquisition time of a single frame?
3. What is the irradiance and the luminance on the sample? Is this below the recommended threshold for the illumination of negatives?
4. In lines 180-181, the paper compares two resolution figures: 700 ppi and 10 line pairs /mm; to make this comparison more readable and understandable, the authors should express both figures with the same units.
5. Figure 2: the image in the pdf file is low-resolution (mostly due to the jpeg compression). In the final publication, high-quality images should be shown. In addition, in the Caption (b) and (c) are switched.
In Panel (a) of the same figure, characters are too small.
6. The authors mention that reference spectral databases of photographic dyes emulsions are not yet available. In principle, with the current HSI system, it would be possible to collect reference spectra and build a database for data analysis. I wander why, in the present study, the authors have not collected a set of reference spectra before analyzing the measured damaged films: is it because no such films are available? The authors should comment on this aspect.
7. Minor comments:
* Line 256: to selectively produce one of the three primary (of is missing)
* Figure 5: in the main text, the authors mention that panel 5d is the PC4 image, while the caption mentions PC3.
* Ref. 44: http://www.ifac.cnr.it/images/stories/libri/archivio/BOOK/Memoria-fotografica.pdft
The extension of the file is .pdf and not pdft, please change that. In addition, this document is in Italian. Is there an English version?
* Refs. 39-43 should be checked.
I believe that the paper should be published after these points have been addressed.
Author Response
Authors acknowledge the Reviewer for the constructive report and gratefully thank for the interest expressed in the proposed prototype.
Please see the attachment for the Reply point to point and the detailed list of revisions.
Author Response File: Author Response.pdf
Reviewer 3 Report
his manuscript provides a novel Transmittance Vis-NIR Hyper-Spectral Imaging scanner for analysis of photographic negatives. This topic is interesting and meaningful. I would like to recommond to accept it after majorrevison. Several specific comments are listed below.
1. It is lack of data analysis.
2. The results are not clearly presented.
3. The format of references should be modified.
4. Discussion should be added to compare with other similar studies.
Author Response
Authors acknowledge the Reviewer for the comments and teh interest shown in the proposed device. The Reply point to point is provided in attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The authors have raised all the suggestions made and the manuscript has been enhanced substantially. Although I sincerely think that the authors should wait until the methodology was fully proved, if the Editors agrees, I think it may be published as it is. Congratulations.Reviewer 3 Report
I would like to accept it in current form.
