Oral and Dental Spectral Image Database—ODSI-DB
Round 1
Reviewer 1 Report
The authors collected a publicly available oral and dental reflectance spectral image database, which contains 316 spectral images from 30 human test subjects. 215 of them were annotated by experts. This dataset may provide some value for some applications such as segmentation.
This paper is well-written. I have some concerns about other potential applications.
- It seems that the main annotation from experts is the segmentation mask, which means that this dataset can be used for image segmentation tasks. I am wondering if the authors could elaborate more on other applications. It would be better to have such a discussion with appropriate references.
- I am wondering how many views were being taken from oral and if this dataset can be used for reconstruction.
- Will the patients-related information be released along with the dataset? For example, sex and age. This may be helpful for some computer vision tasks.
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
First of all, it has been really interesting reading your article, but I have some comments. Maybe you should know that my field of expertise is image, not dentistry, and this can help you to contextualize my comments:
First of all, I miss the presence of specialist in photography or at least in photographic archives in a research about an image database to design the shooting and the design of the database, of the methodology in its creation or of the protocols followed for the photographic taking (types of planes, angles, tools to hold the mouth ...) or the data collection sheet. Taking into account that the base is for images, I understand that there is enough information on dental issues but I miss important information about the shooting speed, whether they have been edited or not, whether they have been masked or not... You label makeup: why if you are photographing to research and you are interested in skin colour... admit makeup?
The use of the mask for the dust particles (too much dust for a lens) can affect some aspects of the image. Why the lens was not cleaned or used another camera? If this is a research based on images: the most important thing should be a perfect lens.
You use sRGB colour but you say It's necessary more information about colour: could you explain or justify why you use this colour space?
If polarizers are important: you should use them and have the tools. If they are not essential, you should explain better that part.
I have the impression that you loss the focus of the introduction of the database as a tool for research and sometimes, especially in the beginning, the article seems to be something different: an analysis of the Spectral Image in front of the other kind of images. You share images but we cannot see the Database appearance. How is the database? Design? Why is important this database? Is the only one? Which model of database have you used? Software or accessibility in the future...? I cannot access: The website explains I have to download it and I cannot do it. It’s necessary to download the hole database? It’s just a zip with images and information? These are the kind of specifications I miss in an article about a database.
The categorization: grayscale, colour and spectral images should be revised as spectral images are also grayscales images and X-ray imaging is not always exactly grayscale. I would recommend to re-examine the tags or names of the three categories.
Figure 3-d: the shooting could be better if the mouth was more open, just one hand is not enough to do so for the documentation of the hole upper teeth. Maybe you need a special tool or It's impossible: you should explain.
Figure 8: the colour labelling appearance is precarious.
Conclusions: the Database needs to be valued. The current conclusion paragraph could be the opening paragraph of the article, but the text needs something more as a conclusion.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
The article describes the database of medical images and its creation. The database itself can be very useful but I think that this type of article is not suitable for journal with high impact factor.
Could you describe similar medical databases (from another field that dentistry) in introduction?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
Thank you for the opportunity to review the updated article.
The article would be more interesting if it also contained a specific case study with a database. The article is good, but I have doubts whether the given type of article (database description) is suitable for publication in a journal with high impact factor. I am missing a description of the specific application where the database was used.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
