Computer-implemented method, computer program, systems and x-ray facility for correction of x-ray image data with regard to noise effects

A method for correction of x-ray image data with regard to noise effects, the method comprising:

• inputting at least one x-ray image dataset;
• determining a noise effect dataset that describes the noise effects using a statistical physics model parameterized by model parameters;
• and
• correcting the x-ray image dataset with the noise effect dataset;
• wherein the model parameters are determined by a trained determination algorithm that inputs the x-ray image dataset and determines the noise effect dataset by applying the model parameters determined by applying the model parameters to the statistical physics model.

A provision system for providing a trained determination algorithm for carrying out a correction process of x-ray image data with regard to noise effects wherein a trained determination algorithm determines, from input data comprising a recorded x-ray image dataset, a noise effect dataset describing the noise effects that is used for correction of the x-ray image data, the provision system comprising:

• a first interface configured to receive one or more training datasets that each comprise x-ray image submasters with assigned known noise effect sub datasets, at least one assigned, noise-free reference image sub dataset, or assigned known noise effect sub datasets and at least one assigned, noise-free reference image sub dataset;
• a computation unit configured to train the determination algorithm using the one or more training datasets, wherein a statistical physics model parameterized by model parameters is used to describe the noise effects, wherein the model parameters are determined at least partly using the determination algorithm;
• and
• a second interface configured to provide the trained determination algorithm.

A correction system for correction of x-ray image data with regard to noise effects, the correction system comprising:

• a first interface configured to receive an x-ray image dataset;
• a correction unit configured to determine, using the x-ray image dataset as input data, a noise effect dataset describing the noise effects, the correction unit further configured to correct the x-ray image dataset using the noise effect dataset, wherein a statistical physics model parameterized by model parameters is used for describing the noise effects, wherein a trained determination algorithm is stored in the correction unit, the trained determination algorithm configured to use the x-ray image dataset as input data in order to output associated model parameters, wherein the correction unit is further configured to determine the noise effect dataset by applying the model parameters determined to the statistical physics model;
• and
• a second interface configured to output a correction result.