Image-based motion compensation of image data

摘要

A method, a computer program, a computer program product and a computed tomography system are disclosed. The image data is a spatially three-dimensional reconstruction. At least one value for an image metric of the image data is determined. A motion field for motion compensation of the image data is then determined on the basis of image data as a function of the image metric. Essentially, partial image data is determined, wherein the partial image data corresponds in each case to the spatially three-dimensional reconstruction from scan data of an angular sub-range. The motion field of the image data is determined at the control points via an optimization method as a function of the image metric, so that, thereafter, the partial image data is transformed in accordance with the motion of the motion field. New image data is then produced by merging the partial image data.

主权项

1. A method for motion compensation of image data of a moving scanned object, the image data being a spatially three-dimensional reconstruction from scan data and the scan data being acquired during a relative rotational movement between a radiation source and the scanned object within an angular range, wherein the angular range includes a plurality of angular sub-ranges, each angular sub-range being assigned to a temporal control point, said method comprising:
(i) determining at least one value for an image metric of the image data, the image metric being a measure of motion artifacts in the image data; (ii) determining partial image data, the respective partial image data corresponding to respective spatially three-dimensional reconstruction from scan data of an angular sub-range; (iii) determining a respective motion field of the image data at a respective control point via an optimization method as a function of the image metric, wherein the respective motion field is a measure of respective motion of the image data; (iv) transforming the respective partial image data according to the motion described by the respective motion field; and (v) generating image data by merging the transformed respective partial image data, wherein the (i) to (v) are performed iteratively until a stop criterion is reached.