A. Goshtasby, S. Nambala, W. de Rijk, S. Campbell
This work was funded by the National Institute of Dental Research.
IEEE Trans. Medical Imaging, vol. 16, no. 5, Oct. 1997, pp. 664-674.
A range scanner is developed that can scan a gypsum dental cast and reconstruct the cast digitally for display and storage purposes. The scanner is based on subtractive light and computes the range values using optical triangulation. A fiducial marker is introduced that, when attached to a dental cast at the time of image acquisition, makes it possible to integrate multiview range images of the cast without image registration. A method for calibrating the scanner is described and experimental results showing the accuracy of the scanner are presented.
Fig. 1. Organization of the scanner.
Fig. 2. (a), (b) Examples of intensity images that have to be processed to obtain
the range images.
Fig. 3. Optical triangulation in the traditional (left) and in this
Fig. 4. (a), (b) Two consecutive image frames. (c) Difference image of the frames.
(d) Zero-crossings of the difference image.
Fig. 5. Vertical stage used to calibrate the scanner (left),
and parameters a, b, and c (right).
Fig. 6. Relation between the camera and the fiducial (left),
and image of the fiducial in the image plane (right).
Fig. 7. Use of the fiducial when scanning a dental
cast and a single tooth.
Fig. 8. Cubes of sides 10mm, 20mm, 30mm, and 40mm that
were used to measure the accuracy of the scanner.
(c) (d) (e)
Fig. 9. (a), (b) Two views of a model tooth. (c) - (g) Scans
from different views
of the tooth. (h) Integration of range images (c)-(g). (i), (j) Two views of
the reconstructed tooth obtained by fitting an elastic surface to (h).
Fig. 10. (a) A mandibular dental cast. (d) - (d) Three
different views of the reconstructed cast.
[WSU Home Page] [CSE Department Home Page] [Intelligent Systems Lab Page]
For more information please contact: A. Goshtasby (firstname.lastname@example.org).