Method and system for the classification of materials by means of machine learning

A classification unit for automatic classification of materials in a spectral medical image recording of an object, comprising:

• a memory storing program computer-readable instructions;
• and
• one or more processors configured to execute the instructions such that the one or more processors are configured to,
• recognize graphic elements in image recordings via an algorithm;
• provision a reference recording set including spectral reference recordings of the object, materials to be classified being annotated in the spectral reference recordings of the object;
• and
• train the classification unit according to at least one principle of machine learning, for the automatic classification of materials in the spectral medical image recording of the object, using the materials annotated in the spectral reference recordings.

A learning computing device, comprising:

• a processor;
• and
• a data storage device storing program computer-readable instructions which, during execution, enable the processor to:
• record reference recordings provided as image data of an object,
• recognize materials in the reference recordings, the reference recordings including spectral reference recordings, materials to be classified being annotated in spectral reference recordings of the object, and graphic elements in the image data being recognizable via an algorithm, and
• train a classification unit according to at least one principle of machine learning, for automatic classification of materials in spectral medical image recording of the object, using the materials annotated in the spectral reference recordings.

A control device for control of a medical imaging system, designed for performance of an automatic classification method of materials in a spectral medical image recording of an object, comprising:

• a processor;
• and
• a data storage device storing program computer-readable instructions which, during execution, enable the processor to
• provision or produce a classification unit trained via a method including
• recognizing graphic elements in image recordings via an algorithm,
• provisioning a reference recording set including spectral reference recordings of the object, materials to be classified being annotated in the spectral reference recordings of the object, and
• training the classification unit according to at least one principle of machine learning, for the automatic classification of materials in the spectral medical image recording of the object, using the materials annotated in the spectral reference recordings;
• provisioning a spectral medical image recording of an object;
• classifying materials in the spectral medical image recording via the classification unit;
• and
• identifying at least one of the materials classified and material-specific processing of the spectral medical image recording.