A description for computed tomography based on sinusoidal curves

Article type: Research Article

Authors: Liu, Julian J.^; | Watt-Smith, Stephen R. | Smith, Stephen M.

Affiliations: Oxford Centre for Functional Magnetic Resonance Imaging of the Brain, University of Oxford, John Radcliffe Hospital, Oxford, OX3 9DU, UK. Tel.: +44 186 522 2739; Fax: +44 186 522 2717; E-mail: liu@fmrib.ox.ac.uk | Nuffield Department of Surgery, University of Oxford, UK

Abstract: A new description for computed tomography, parallel and equal to Radon transformation but based on sinusoidal functions rather than line integrals, is introduced. This representation contributes an effective algorithm to reduce metal artefacts in reconstruction. Using this method, the information included in the scan data corresponding to the metal implants can be separated from the original projection, an amending method rather than interpolation is therefore provided to reach higher accuracy for reconstruction. The method is with low complexity in computation, and can be combined online with filtered backprojection (FBP), which is the most widely used algorithm in practice, to improve the quality of the reconstructed image. Also, the method has the potential to deal with the artefacts caused by beam hardening and partial volume, and to be developed into a straight reconstruction algorithm based on the sinusoidal representation. Examples are presented for clearer description and demonstration.

Journal: Journal of X-Ray Science and Technology, vol. 11, no. 4, pp. 205-218, 2003