ODSI-DB – Oral and Dental Spectral Image Database
Creators
-
1.
University of Eastern Finland
Description
ODSI-DB is an image database of oral and dental reflectance spectral images of human test subjects. Image sets of the test subjects contain the front-view and the occlusal surfaces of lower and upper teeth, oral mucosa, and face surrounding the mouth. Other features-of-interest have been imaged on case-by-case basis. The spectral images in the database have been annotated by dental experts.
Technical info (English)
Oral and Dental spectral images
The oral and dental spectral image files are multipage TIFF images in dataset.7z. The first page contains a 24-bit RGB preview image, followed by grayscale spectral band images (16-bit grayscale). The files utilize tag system of TIFF files to store the spectral band center wavelengths in tag 65000 as list of 32-bit floating point values, and tag 65111 to store a free-form ASCII metadata string. The metadata string contains information about the imaging setup, image processing and miscellaneous measurement data (blood pressure and oxygen saturation before and after measurement). File dataset-metadatas.txt contains these free-form metadata strings per spectral image.
Annotation files
Annotation files are likewise multipage TIFF images. Each page of the annotation image file contains a binary black-and-white mask image and has the annotation class stored as an ASCII-string in tag 65001. File dataset-labels.txt contains annotation classes and numbers of pixels per class per mask image.
Software
Python functions for reading and writing spectral image TIFFs and annotation mask TIFFs are provided in the spectral_tiffs.py file in the software.7z archive.
Other
The complete uncompressed spectral image database is 34 GB in size.
For further questions or any issues regarding the ODSI-DB database, please contact the Spectral Color Research group, Professor Markku Hauta-Kasari, by email markku.hauta-kasari@uef.fi.
Notes (English)
Funding
Acknowledgments
Files
dataset-labels.txt
Additional details
Related works
- Is published in
- Journal article: 10.3390/app10207246 (DOI)
- Is supplemented by
- Software: 10.5281/zenodo.8046363 (DOI)
Funding
- Spectral sensor technology and digital spectral image databases for oral and dental applications (DIGIDENT) 4465/31/2017
- Business Finland
- Photonics Research and Innovation / Consortium: PREIN 320166
- Research Council of Finland