Reduction of Half-Scan Shading Artifact Based on Full-Scan Correction 1

Abstract

Temporal studies (such as blood perfusion) commonly are used to measure function. Radiation dosage is a primary limiting factor for these scans. Half-scan reconstruction can decrease dosage and improve temporal resolution, but is not viable for quantitative studies because of shading artifact. We propose a method for identifying the artifact and minimizing its effect. It is possible to measure the shading artifact by producing both a full-scan and a half-scan reconstruction from the same projection data. A correlation was shown between the subset of data used for reconstruction and per-pixel variation. Furthermore, this variation can be parameterized by only the center angle of the projection data. By performing a single full-scan acquisition, it is possible to generate many half-scan reconstructions and measure the artifact; then future half-scan acquisitions can be corrected. The artifact is caused by the inhomogeneity in the object being scanned. Before correction, the root mean square error between the half-scan reconstruction and the full-scan is 41.0. After correction, the error is decreased to 10.7, or 26% of the original value. We present a method that can measure and correct for object-dependent half-scan shading artifact. This can enable half-scan reconstruction for use in quantitative temporal studies.