A New CT Rawdata Redundancy Measure applied to Automated Misalignment Correction

In computed tomography, redundantly measured rays (i.e. multiple measurements of a line integral through an object) pose a computationally efficient possibility to quantify the rawdata quality in a cost function and thus to reduce many kinds of artifacts. A general downside of such a cost function is that the portion of redundant rays is generally small and depends on the specific data acquisition geometry. The authors propose using information associated to plane integrals instead of line integrals in order to tremendously increase the rawdata utilization when formulating a cost function that quantifies the rawdata quality. In the cone–beam data acquisition geometry, which does not allow plane integrals to be measured directly, plane–information is obtained using the 2D Radon transform of the measured rawdata and a subsequent differentiation operation. The new rawdata redundancy measure is successfully applied for automatic misalignment correction.