Share Email Print
cover

Proceedings Paper

Cone beam reconstruction algorithm for PET-VCT
Author(s): J. Hsieh; A. Lonn; M. Nyka
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

A typical X-ray CT scanner has a 50cm reconstruction field-of-view (FOV) which is significantly smaller than the 70cm FOV typically used for PET. When CT images are used for attenuation correction in PET, significant artifacts result when portions of the scanned object extend beyond 50 cm. In this paper, we present a reconstruction algorithm to estimate the missing projection samples so that a significantly larger reconstruction FOV can be achieved. The estimation is based on the observation that the total attenuation is a continuous function over views. The initial estimate of the projection distribution makes use of the property that the projection profile does not change abruptly as a function of the projection angle. When combined with the projection values at the location of truncation, the size and location of water cylinders that best fit these parameters are calculated. Additional adjustments are made to ensure the total attenuation of the expanded projection samples matches closely to the estimated total missing projection. The estimated projection is further blended with the mirror image of the measured projections near the location of truncation. Various phantom experiments were conducted to evaluate the proposed algorithm. Both axial and helical scans were collected on a 64-slice scanner. Results show that the proposed algorithm can restore the fidelity of the reconstruction for the portion of the object inside 50 cm FOV, and provide an adequate impression of the object outside 50cm.

Paper Details

Date Published: 7 September 2006
PDF: 9 pages
Proc. SPIE 6318, Developments in X-Ray Tomography V, 63180H (7 September 2006); doi: 10.1117/12.681209
Show Author Affiliations
J. Hsieh, GE Healthcare Technologies (United States)
A. Lonn, GE Healthcare Technologies (United States)
M. Nyka, GE Healthcare Technologies (United States)


Published in SPIE Proceedings Vol. 6318:
Developments in X-Ray Tomography V
Ulrich Bonse, Editor(s)

© SPIE. Terms of Use
Back to Top