Automatic failure detection in magnetic resonance apparatuses

1. A method for automatic failure analysis of a medical imaging apparatus during operation of the medical imaging apparatus, comprising:

reading in input data, comprising raw data or image data, acquired by the medical imaging apparatus into a computer; in said computer, calculating a set of performance indicators in the input data, the set of performance indicators comprising: an average signal per channel of said medical imaging apparatus, a highest relative Fast Fourier Transform (FF1) peak, and an average signal-to-noise ratio (SNR);

from said computer, accessing a trained neural network system with said calculated performance indicators in order to use said trained neural network system to produce result data such that, if a failure of said medical imaging apparatus exists, a cause of said failure of said medical imaging apparatus is identified, and otherwise an operational state of the medical imaging apparatus is identified; and from said computer, presenting said result data at a display screen.

2. A method as claimed in claim 1 comprising, in said computer, organizing said result data for display at said display screen so as to classify multiple possible failure causes of the medical imaging apparatus with respect to a probability measure of each of said multiple failure causes.

3. A method as claimed in claim 1 comprising calculating said performance indicators in said computer by automatically directly extracting said performance indicators from raw data in said image data, before reconstruction of image data.

4. A method as claimed in claim 1 comprising calculating said performance indicators in said computer by automatically executing a performance measurement algorithm that operates on said input data.

5. A method as claimed in claim 1 further comprising:

in said computer, detecting changes of said performance indicators over time; and

accessing said trained neural network system with said detected changes of said performance indicators.

6. A method as claimed in claim 1 comprising training said neural network so as to learn a cause of failure of the medical imaging apparatus for a set of performance indicators or changes in said performance indicators over time.

7. A method as claimed in claim 1 comprising reading into said computer, with said input data, data selected from the group consisting of text data, stream data, historic data, and parameter data.

8. A method as claimed in claim 1 comprising incrementally retraining said trained neural network based on an evaluation of said result data.

9. An automatic failure analyzer for use with a medical imaging apparatus, the automatic failure analyzer comprising:

an input interface configured to read in input data, comprising raw data or image data, acquired by the medical imaging apparatus;

one or more processors configured to calculate a set of performance indicators in the input data, the set of performance indicators comprising an average signal per channel of said medical imaging apparatus, a highest relative Fast Fourier Transform (FF1′) peak, and an average signal-to-noise ratio (SNR);

a trained neural network system configured to produce result data as a correlation between performance indicators and causes of failure of said medical imaging apparatus; and

an output interface configured to present said result data at a display screen.

10. The medical imaging apparatus of claim 9 , further comprising:

a storage device configured to store at least one of the trained neural network system and a second trained neural network,

wherein the second trained neural network is configured to learn further sets of performance indicators.