RSNA 2003 

Abstract Archives of the RSNA, 2003


K21-1042

Windmill Artifacts in 16-slice CT

Scientific Papers

Presented on December 3, 2003
Presented as part of K21: Physics (CT: Image Quality Dose)

Participants

Efrat Shefer PhD, PRESENTER: Nothing to Disclose

Abstract: HTML Purpose: 16-slice CT suffers from windmill artifacts in helical scans due to under sampling along the axis of rotation in the presence of partial volume data of a scanned object. The windmill artifact appears as black and white pattern rotating around high contrast features that change along the longitudinal axis. In this work we study the dependence of the windmill artifact on the scan and reconstruction parameters. Methods and Materials: A phantom containing an 80 mm thick, 5 mm diameter metal disc perpendicular to the axis of rotation was used for this study. The frequency of the white and black pattern was measured on the axial images as a function of the pitch and of the number of slices. The windmill artifact level was quantified on MPR images using mean square differences method. The dependence of the artifact level on the pitch, the distance from the center of rotation, and on the number of slices was investigated. A comparison of the windmill artifacts in true 3D Cone Beam and standard 2D back-projection schemes was performed. A schematic model explaining the observed differences has been developed. Results: The set of measured data was reconstructed in standard 2D (fan beam) and true 3D (cone beam) reconstruction algorithms. The dependence of the windmill artifact level on the scan pitch at the center of rotation is similar for 2D and 3D back projection and is a wave-like function. The windmill artifact level at varying distances from the center of rotation was measured. It is shown that at large distance from the center of rotation the artifact level is higher for 2D compared to 3D back projection. The windmill artifact level is reduced for large z-filtering broadening factor used for spiral interpolation. The Frequency of the windmill pattern increases with the pitch and depends on the number of slices. Some of the parameter-combinations were qualitatively verified on a head phantom with a cast human skull. Conclusion: The windmill artifact level and frequency depend strongly on the pitch and on the distance from the center of rotation. At a large distance from the center of rotation the windmill artifact is reduced using true 3D back projection compared to 2D back projection. It is therefore possible to minimize windmill artifacts by optimizing the scan and reconstruction parameters. (E.S., G.S., A.A., D.B. are employees of Philips Medical Systems & Technologies.) Questions about this event email: efrat.shefer@philips.com

Cite This Abstract

Shefer PhD, E, Windmill Artifacts in 16-slice CT.  Radiological Society of North America 2003 Scientific Assembly and Annual Meeting, November 30 - December 5, 2003 ,Chicago IL. http://archive.rsna.org/2003/3104065.html