Share Email Print

Proceedings Paper

Surface reconstruction from structured-light images for radiation therapy
Author(s): Olesya Peshko; Christopher Kumar Anand; Tamás Terlaky
Format Member Price Non-Member Price
PDF $17.00 $21.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

To design and deliver proper radiation treatment for cancer patients, knowledge of the body's surface in the affected area is required. Currently, surface information is obtained by using a manually operated tracer. The drawbacks of this contact method include slow operation, and errors in repositioning the patient in an x-ray machine. Utilization of MRI or CT is also possible but expensive. We propose a non-contact, quick, inexpensive method to reconstruct the surface. In our non-contact method, a mask with transparent circular coloured spots and a black background, and an incoherent light source are used to create structured-light images. Colour coding is necessary to establish the correspondence between the projected and the observed patterns, which is essential for surface reconstruction. The deformed light pattern is photographed by an offset camera and analyzed. First, noise reduction is performed because images are noisy due to the low-light conditions and low sensitivity of an off-the-shelf camera. Then, pattern elements (light elliptical spots) are found in the image. We use an inverse polynomial to model the intensity of a light spot, which results in a non-convex, least-squares optimization problem. Next, spots are assigned to a grid according to their colours and location, and errors are corrected using the relative position of the spots. Finally, spatial coordinates of the surface points are computed and surface reconstruction is performed. The described algorithms are implemented as a MATLAB package, which converts the acquired images into a three-dimensional surface. The developed system is inexpensive, and it can easily be mounted on an x-ray machine. The software package can run on any standard PC.

Paper Details

Date Published: 13 October 2005
PDF: 10 pages
Proc. SPIE 5969, Photonic Applications in Biosensing and Imaging, 59691U (13 October 2005); doi: 10.1117/12.628459
Show Author Affiliations
Olesya Peshko, McMaster Univ. (Canada)
Christopher Kumar Anand, McMaster Univ. (Canada)
Tamás Terlaky, McMaster Univ. (Canada)

Published in SPIE Proceedings Vol. 5969:
Photonic Applications in Biosensing and Imaging
Brian C. Wilson; Richard I. Hornsey; Warren C. W. Chan; Ulrich J. Krull; Robert A. Weersink; Kui Yu, Editor(s)

© SPIE. Terms of Use
Back to Top