Loading patent…
AssigneeSiemens Healthcare GmbH
InventorEric FOURNIE
Estimating the term…
A method is for processing parameters of a machine-learning method for providing a correction dataset for motion correction of a CT image dataset of an object that moves during recording thereof. A training dataset with a number of reference image datasets for introduction into the machine-learning method is provided. Iteration is performed. In each of the iteration steps, a correction dataset is determined for each of the reference image datasets via the machine-learning method and a result of a cost function dependent on the correction datasets determined in this iteration step is ascertained. The iteration is terminated or a parameter of the machine-learning method is altered in dependence on the result of the cost function. A reconstruction method further uses the processed machine-learning method.
A method for processing parameters of a machine-learning method for providing a correction dataset for motion correction of a CT image dataset of an object that moves during recording thereof, the method comprising:
A reconstruction method for generating a motion-corrected reconstruction image dataset for a CT image dataset of an object moving during recording, the reconstruction comprising:
An apparatus, comprising:
Dependencies are inferred from the claim text.