Provision of position information of a local RF coil

Abstract

A computer-implemented method for provision of a result dataset having position information of a local radio-frequency coil, including: providing input data having at least magnetic resonance data, which is acquired by means of the local radio-frequency coil; determining a result dataset by applying a trained function to the input data, wherein the result dataset comprises position information for determining the position of the local radio-frequency coil; and providing the result dataset.

Claims

1. A computer-implemented method for providing a result dataset including position information of a local radio-frequency coil, the method comprising: providing input data including input magnetic resonance data, which is acquired by means of the local radio-frequency coil, wherein the input magnetic resonance data is normalized with respect to magnetic resonance data acquired via a radio-frequency body coil by performing a division of magnetic resonance data acquired via the local radio-frequency coil by magnetic resonance data acquired via the radio-frequency body coil for each spatial point, and wherein a receive profile of the magnetic resonance data acquired via the radio-frequency body coil is more homogenous than a receive profile of the magnetic resonance data acquired via the local radio-frequency coil; determining a result dataset by applying a trained function to the input data, the trained function comprising an artificial neural network (ANN), wherein the result dataset includes position information for determining the position of the local radio-frequency coil; and providing the result dataset.

2. The method of claim 1, wherein the input magnetic resonance data includes position information in a spatial direction.

3. The method of claim 1, wherein the input magnetic resonance data includes non-spatially encoded magnetic resonance data with different measurement frequencies.

4. The method of claim 1, wherein the input data includes a further item of coil information of the local radio-frequency coil.

5. The method of claim 1, wherein the result dataset includes a match value, which specifies how likely there is to be a match between the position of the local radio-frequency coil established from the position information and an actual position of the local radio-frequency coil.

6. The method of claim 1, wherein the result dataset is provided for a number of local radio-frequency coils and/or for a number of coil elements of a local radio-frequency coil at the same time.

7. A non-transitory computer-readable memory medium on which program sections which are readable and executable by a provision system are stored, for carrying out the method of claim 1 when the program sections are executed by the provision system.

8. The method of claim 1, wherein the input data comprises magnetic resonance data acquired from different regions of a patient body via respective local training radio-frequency coils.

9. The method of claim 1, wherein the magnetic resonance data of the radio-frequency body coil is acquired separately from the magnetic resonance data acquired via the local radio-frequency coil.

10. A computer-implemented method for providing a trained function, comprising: receiving or determining a training dataset of a local training radio-frequency coil, wherein the training dataset includes a training input dataset and a training result dataset, and the training input dataset includes input magnetic resonance data of the local training radio-frequency coil, wherein the input magnetic resonance data is normalized with respect to magnetic resonance data acquired via a radio-frequency body coil by performing a division of magnetic resonance data acquired via the local radio-frequency coil by magnetic resonance data acquired via the radio-frequency body coil for each spatial point, and wherein a receive profile of the magnetic resonance data acquired via the radio-frequency body coil is more homogenous than a receive profile of the magnetic resonance data acquired via the local radio-frequency coil; determining a result dataset by applying the trained function to the training input dataset, the trained function comprising an artificial neural network (ANN); adapting a parameter of the trained function based on a comparison between the training result dataset and the result dataset; and providing the trained function.

11. The method of claim 10, wherein the trained function is based on the training dataset with training data, and the training data comprises the input magnetic resonance data acquired via different local training radio-frequency coils.

12. The method of claim 10, wherein the trained function is based the training dataset with training data, and the training data comprises the input magnetic resonance data with position information in a spatial direction.

13. The method of claim 10, wherein the trained function is based on the training dataset with training data, and the training data comprises non-spatially-encoded magnetic resonance data with different measurement frequencies.

14. The method of claim 10, wherein the trained function is based on the training dataset with training data, and the training data comprises a coil type and/or an examination region and/or a couch position and/or a position of the patient and/or anatomy information of the patient.

15. The method of claim 10, wherein the trained function comprises at least two hidden layers and a maximum of ten hidden layers.

16. The method of claim 10, wherein the trained function comprises a layer with LSTM (Long Short Term Memory) neurons.

17. The method of claim 10, wherein the trained function comprises a hidden layer embodied as a drop-out layer.

18. A non-transitory computer-readable memory medium on which program sections which are readable and executable by a training system are stored, for carrying out the method of claim 10 when the program sections are executed by the training system.

19. A provision system for provision of a result dataset, comprising: an interface; and a processor, wherein: the interface and/or the processor are configured to provide input data, the processor is configured to determine a result dataset by applying a trained function to the input data including input magnetic resonance data of a local radio-frequency coil, the trained function comprising an artificial neural network (ANN), the input magnetic resonance data is normalized with respect to magnetic resonance data acquired via a radio-frequency body coil by performing a division of magnetic resonance data acquired via the local radio-frequency coil by magnetic resonance data acquired via the radio-frequency body coil for each spatial point, a receive profile of the magnetic resonance data acquired via the radio-frequency body coil is more homogenous than a receive profile of the magnetic resonance data acquired via the local radio-frequency coil, the result dataset includes position information of the local radio-frequency coil, and the interface is further configured to provide the result dataset.

20. A magnetic resonance apparatus, comprising the provision system of claim 19.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Further advantages, features and details of the disclosure emerge from the exemplary aspects described below and also with the aid of the drawings. In the figures:

(2) FIG. 1 shows a method for provision of a result dataset comprising position information of a local radio-frequency coil.

(3) FIG. 2 shows a method for provision of a trained function.

(4) FIG. 3 shows a trained function in a schematic diagram.

(5) FIG. 4 shows a provision system in a schematic diagram.

(6) FIG. 5 a training system in a schematic diagram.

(7) FIG. 6 shows a magnetic resonance apparatus with a provision system in a schematic diagram.

DETAILED DESCRIPTION

(8) Shown in FIG. 1 is a method for provision of a result dataset comprising position information of a local radio-frequency coil 600. In a first method step 100 there is a provision of input data, wherein the input data at least comprises magnetic resonance data, which is acquired by means of the local radio-frequency coil 600. The input data is preferably provided by means of a provision system 400, as is described in the remarks relating to FIG. 4 below. In particular the input data comprising magnetic resonance data acquired at least by means of the local radio-frequency antenna 600 is provided by means of an interface 401, in particular by means of an interface 401 of the provision system 400. The input data has been acquired before the beginning of the method by means of the local radio-frequency coil 600 and is transferred from the local radio-frequency coil 600 to the provision system 400.

(9) In this case the input data can also comprise magnetic resonance data, which is acquired by means of a local radio-frequency coil 600 or also by means of two or more local radio-frequency coils 600 and/or also by means of two or more coil elements of a local radio-frequency coil 600. The input data can moreover also comprise additional information, such as for example which local radio-frequency coil 600 and/or which coil elements of the local radio-frequency coil 600 have been used for the acquisition of magnetic resonance data included in the input data. Moreover the input data can also comprise additional information, such as for example a region to be examined of the patient and/or a couch position and/or position information of the patient and/or anatomy information, in particular size information, of the patient and/or further information appearing sensible to the person skilled in the art.

(10) In this first method step 100 the input data provided here can comprise magnetic resonance data acquired by means of the local radio-frequency coil 600, wherein the magnetic resonance data comprises position information in at least one spatial direction. The position information of the magnetic resonance data in this case can preferably comprise a spatial encoding of the acquired magnetic resonance data. For example the magnetic resonance data with position information, in particular spatial encoding, can be acquired by means of a spin echo measurement and/or a gradient echo measurement. The position information, in particular the spatial encoding, of the magnetic resonance data in this case can comprise one spatial direction or a number of spatial directions. The magnetic resonance data with position information, in particular a spatial encoding, in a number of spatial directions and/or a number of spatial coordinates can be present for each spatial direction and/or each spatial coordinate as separate magnetic resonance data, in particular as a separate spectrum. As an alternative or in addition it is also conceivable for the magnetic resonance data with position information, in particular a spatial encoding, to be present in a number of spatial directions and/or a number of spatial coordinates as multidimensional magnetic resonance data, in particular as a multidimensional spectrum. The magnetic resonance data, in particular the spectra, can further be present already Fourier-transformed or also in the k-space encoding.

(11) As an alternative or in addition, in this first method step, the input data provided here can comprise magnetic resonance data acquired by means of the local radio-frequency coil 600, wherein the magnetic resonance data of the local radio-frequency coil 600 is normalized in relation to magnetic resonance data of a homogeneous radio-frequency coil. The homogeneous radio-frequency coil preferably comprises a body coil. In particular here a receive profile of acquired magnetic resonance data and/or signals of the body coil is more homogeneous than for example a receive profile of acquired magnetic resonance data and/or signals of a knee coil and/or further local radio-frequency coils. A normalization of the acquired magnetic resonance data can for example comprise a division of the magnetic resonance data acquired by means of the local radio-frequency coil 600, in particular magnetic resonance data with a spatial encoding, by the magnetic resonance data and/or signals acquired with the homogeneous radio-frequency coil, in particular the body coil, for each spatial point. The magnetic resonance data and/or signals of the homogeneous radio-frequency coil in this case can also be acquired separately from the acquisition of the magnetic resonance data by means of the local radio-frequency coil 600.

(12) As an alternative or in addition, in this first method step the input data provided here can comprise magnetic resonance data acquired by the local radio-frequency coil 600, wherein the magnetic resonance data comprises non-spatially-encoded magnetic resonance data with different measurement frequencies. The magnetic resonance data can comprise in this case with a measurement frequency, which comprises an average resonant frequency in the volume close to the center of a scanner unit of a magnetic resonance apparatus. Moreover the magnetic resonance data can also comprise measurement frequencies, which deviate from the average resonant frequency. Through this in particular local radio-frequency coils 600 and/or coil elements of a local radio-frequency coil 600 can be identified and/or position information reliably acquired from local radio-frequency coils 600 and/or coil elements of a local radio-frequency coil 600, with said local radio-frequency coils 600 and/or coil elements of a local radio-frequency coil 600 preferably being arranged and/or positioned outside of a homogeneity area and/or linearity area of a scanner unit 602 of a magnetic resonance apparatus 601.

(13) Moreover it can also be that the input data provided, in particular magnetic resonance data acquired by means of the local radio-frequency coil 600, is already pre-processed, such as for example by smoothing and/or filtering of the input data.

(14) Subsequently, in a further, second method step 101, there is a determination of a result dataset by application of a trained function 300 to the input data, wherein the result dataset comprises position information for determining the position of the local radio-frequency coil 600. The result dataset is preferably determined by means of a determination unit and/or a processing unit 402, in particular by means of a determination unit and/or a processing unit 402 of the provision system 400.

(15) The trained function 300 preferably comprises an artificial neural network. In this way the provision of a result dataset comprising position information of a local radio-frequency coil 600 is based in particular on a machine learning process, which is based on the artificial neural network. An artificial neural network (ANN) is in particular a network of artificial neurons emulated in a computer program. The artificial neural network in this case is typically based on a networking of a number of artificial neurons. The artificial neurons in this case are typically arranged on different layers. Usually the artificial neural network comprises an input layer 301 and an output layer 303, of which the neuron output is visible as the only layer of the artificial neural network. Layers lying between the input layer and the output layer are typically referred to as hidden layers 302. Further information about the trained function 300 and/or of the artificial neural network is provided below in the explanation of FIG. 3.

(16) The trained function 300 and/or the artificial neural network has in particular already been suitably trained in advance for the determination of position information for determining the position of the local radio-frequency coil 600 with the aid of the acquired magnetic resonance data. Training datasets are used in particular in this case for the training of the trained function 300 and/or of the artificial neural network, in which for example a signal intensity of the magnetic resonance data acquired by means of the local radio-frequency coil is assigned to a position and/or position information of the local radio-frequency coil. The medical training datasets in this case are typically acquired from training persons and/or training radio-frequency coils different from the patient.

(17) The result dataset preferably comprises the position information for determining the position of the local radio-frequency coil 600 and/or a position of coil elements of a local radio-frequency coil 600. The result dataset in this case can also directly comprise the position of the local radio-frequency coil 600 and/or the position of coil elements of a local radio-frequency coil 600. Moreover the result dataset can also comprise further information relating to the position of the local radio-frequency coil 600 and/or relating to the position of coil elements of a local radio-frequency coil 600. For example the result dataset here can also comprise information and/or a match value, which specify how likely it is that there is a match between the position of the local radio-frequency coil 600 and/or of coil elements of a local radio-frequency coil 600 established from the position information and the actual position of the local radio-frequency coil 600 and/or of coil elements of a local radio-frequency coil 600. Moreover it can also be that the result dataset also contains information that no local radio-frequency coil 600 and/or no coil element of a local radio-frequency coil 600 could be determined or that a position could not be determined for any local radio-frequency coil 600 and/or any coil element of a local radio-frequency coil 600. In such a case the position information can assume the value “0” for example.

(18) In a further, third method step 102 following on from this there is a provision of the result dataset. The result dataset is preferably provided by means of the provision system 400, in particular by means of the interface 401 of the provision system 400. In this case, in this third method step 102, the result dataset can also be provided for a number of local radio-frequency coils 600 and/or for a number of coil elements of a local radio-frequency coil 600 at the same time. Preferably the input data provided here also comprises the information as to the local radio-frequency coils 600 and/or coil elements of a local radio-frequency coil 600 for which magnetic resonance data is available. In this case, when different combinations of local radio-frequency coils 600 and/or of coil elements of a local radio-frequency coil 600 are used, different parameterizations of the trained function 300 and/or of the artificial neural network can be used.

(19) If for example for a measurement the local radio-frequency coils 600 of type A and type B are used together, this data can be passed together in the first method step 101 to the trained function 300 and/or to the artificial neural network. Thus a different parameterization of the trained function 300 and/or of the artificial neural network is used than if local radio-frequency coils 600 of type A and C or just one local radio-frequency coil 600 of type A were used on its own.

(20) Shown in FIG. 2 is a computer-implemented method for provision of the trained function 300, wherein the trained function 300 is embodied to determine a result dataset with the aid of input data. The trained function 300 is preferably provided by means of a training system 500.

(21) In a first method step 200 of the method for provision of the trained function 300 there is a receipt or determination of at least one training dataset of a local training radio-frequency coil, wherein the at least one training dataset comprises a training input dataset and a training result dataset and the training input dataset comprises magnetic resonance data of the local training radio-frequency coil. The at least one training dataset, in particular the training input dataset of the at least one training dataset, in particular has magnetic resonance data, which is acquired for example by means of the local training radio-frequency coil. Moreover it can also be that for training of the trained function 300 and/or of the artificial neural network also, instead of real magnetic resonance data, simulated magnetic resonance data is used. For example a dataset simulated in this way, in particular a simulated magnetic resonance dataset, of a local training radio-frequency coil can be provided by means of a Bloch simulation and/or by further simulation methods appearing sensible to the person skilled in the art.

(22) The at least one training dataset is received or determined in particular by means of a training processing unit 502 and/or a training interface 501, in particular by means of the training processing unit 502 and/or the training interface 501 of the training system 500 (see FIG. 5). The result dataset is determined and the parameters of the trained function 300 and/or of the artificial neural network adapted in particular by means of the training processing unit 502 of the training system 500. The trained function 300 and/or the artificial neural network are provided in particular by means of the training interface 501 of the training system 500.

(23) Preferably, for a provision of the trained function 300 and/or of an artificial neural network, training datasets, in particular the training input datasets and training result datasets of different local training radio-frequency coils and/or of different coil elements of a local training radio-frequency coil are made available. Moreover the training datasets, in particular training input datasets and training result datasets, comprise training data of coil combinations of a number of local training radio-frequency coils and/or of a number of coil elements of a local training radio-frequency coil. Moreover, for the provision of the trained function 300 and/or of an artificial neural network, training datasets, in particular training input datasets and training result datasets, can be made available, which were acquired from different regions of the body by means of a local training radio-frequency coil and/or by coil elements of a local training radio-frequency coil.

(24) In this first method step 200 the at least one training dataset, in particular the training input dataset of the at least one training dataset, can comprise training data, wherein the training data comprises magnetic resonance data with position information in at least one spatial direction. The position information, in particular the spatial encoding, of the training data can be undertaken in this case in one spatial direction or in a number of spatial directions. In this case, for each spatial direction and/or spatial coordinate, separate training data, in particular a separate training spectrum, can be present. Moreover it is also conceivable for multidimensional training data, in particular a multidimensional training spectrum, to be present here. The training data, in particular the training spectra, can further be present here already Fourier-transformed or also in the k-space encoding.

(25) Furthermore, in this first method step 200, the at least one training dataset, in particular the training input dataset of the at least one training dataset, can comprise training data, wherein the training data comprises non spatially-encoded magnetic resonance data with different measurement frequencies. The training data can in this case comprise with a measurement frequency, which comprises an average resonant frequency in the volume close to the center of the scanner unit 602 of the magnetic resonance apparatus 600. Moreover the training data can also comprise measurement frequencies, which deviate from the average resonant frequency.

(26) Furthermore, in this first method step 200, the at least one training dataset, in particular the training input dataset of the at least one training dataset, can comprise training data, wherein the training data comprises further coil information, in particular a coil type and/or an examination region and/or a couch position. Moreover in this first method step 200, the training data can also comprise a position and/or position information of the patient and/or anatomy information of the patient. The anatomy information can for example comprise a size and/or extent of the patient.

(27) In a subsequent second method step 201 of the method for provision of the trained function 300 there is a determination of a result dataset by application of the trained function 300 to at least one training dataset, in particular to the training input dataset. The trained function 300 and/or the artificial neural network preferably comprises an input layer, a number of hidden layers and an output layer, as is shown in FIG. 3. Preferably the trained function 300 and/or the artificial neural network comprises at least two hidden layers and a maximum of ten hidden layers. Preferably the trained function 300 and/or the artificial neural network comprises at least two hidden layers and a maximum of eight hidden layers. Especially advantageously the trained function 300 and/or the artificial neural network comprises at least three hidden layers and a maximum of five hidden layers.

(28) The trained function 300 and/or the artificial neural network can also comprise a fully connected neural net in this case, in which each neuron of a layer is connected to each neuron of the preceding layer and of the succeeding layer. Moreover the trained function 300 and/or the artificial neural network can comprise at least one layer with LSTM neurons (Long Short Term Memory neurons). Here there can be feedback between the neurons of different layers. This variant of the trained function 300 and/or of the artificial neural network above all comprises an effective learning phase, in that in multilayer pure feed-forward networks in particular, i.e. in multilayer networks without feedback, the problem of parameters and/or weights of the front hidden layers only being inadequately optimized during the learning phase, can be reduced and/or prevented.

(29) The trained function 300 and/or the artificial neural network can comprise at least one hidden layer embodied as a drop-out layer. Such drop-out layers comprise a regularization method in order to reduce and/or to prevent an overfitting of the trained function 300 and/or of the artificial neural network. Here, during the training of the trained function 300 and/or of the artificial neural network, individual neurons in the drop-out layers chosen at random are deactivated and not taken into account for the next computation step.

(30) In a subsequent third method step 202 of the method for provision of the trained function 300, at least one parameter of the trained function 300 is adapted based on a comparison of the training result dataset of the at least one training dataset and the result dataset. In this third method step 202 parameters and/or weights of the links between two neurons of the trained function 300 and/or of the artificial neural network are defined. In particular the parameters and/or weights of the links between two neurons of the trained function 300 and/or of the artificial neural network are defined by means of supervised learning with back propagation. The parameters and/or weights are thus optimized by means of backpropagation. The training of the trained function 300 and/or of the artificial neural network should take place individually where possible for each coil type of a local radio-frequency coil, possibly also for widely-used coil combinations of local radio-frequency coils in order to define suitable parameters and/or weights or neuron connections for each local radio-frequency coil or for each widely-used coil combination of local radio-frequency coils.

(31) In a subsequent fourth method step 203 of the method for provision of the trained function 300 the trained function 300 is provided. The trained function 300 and/or the artificial neural network are preferably provided by means of the training interface 501 of the training system 500. The provision can in particular comprise a storage, display and/or transmission of the trained function 300 and/or of the artificial neural network. In particular, the trained function 300 and/or the artificial neural network can be transmitted to the provision system 400 or used in a method for provision of a result dataset in accordance with the disclosure and its aspects.

(32) Shown in FIG. 3 in greater detail is the trained function 300 and/or the artificial neural network. The trained function 300 and/or the artificial neural network comprises an input layer 301, a number of hidden layers 302 and an output layer 303.

(33) The input layer 301 in this case can comprise input data 304, which comprises magnetic resonance data with position information in at least one spatial direction. Furthermore the input layer 301 can comprise input data 305, which comprises non-spatially-encoded magnetic resonance data, which comprises a measurement frequency that is the same as the resonant frequency. The input layer 301 can furthermore comprise input data 306 with non-spatially-encoded magnetic resonance data, which comprises a measurement frequency that is different from the resonant frequency. Furthermore the input layer 301 has further input data 307, which comprises additional coil information.

(34) The hidden layers 302 preferably comprise at least two hidden layers 302 and a maximum of ten hidden layers 302. Preferably the trained function comprises at least two hidden layers 302 and a maximum of eight hidden layers 302. Especially advantageously the trained function comprises at least three hidden layers 302 and a maximum of five hidden layers 302. The hidden layers are only shown schematically in FIG. 3 and do not reproduce the links of the individual neurons between the individual layers as described above.

(35) The output layer 303 comprises the result dataset provided. This result dataset comprises the position information 308 and the match value 309.

(36) Shown schematically in FIG. 4 is the provision system 400. The provision system 400 shown is embodied to carry out a disclosed method for provision of a result dataset comprising position information of a local radio-frequency coil 600 and/or of coil elements of a local radio-frequency coil 300. The provision system 400 comprises an interface 401, a processor unit 402 and a memory unit 403.

(37) The provision system 400 can in particular involve a computer, a microcontroller or an integrated circuit. As an alternative the provision system 400 can involve a real or virtual group of computers (a real group is a cluster and a virtual group is a cloud). The provision system 400 can also be embodied as a virtual system, which is executed on a real computer or on a real or virtual group of computers (the technical term is virtualization).

(38) The interface 401 can involve a hardware interface or software interface (for example PCI bus, USB or Firewire). The processor unit 402 can have hardware elements or software elements, for example a microprocessor or what is known as an FPGA (acronym for Field Programmable Gate Array). Moreover the processor unit 402 can comprise components that are specialized and/or optimized for tasks and/or applications for machine learning, such as for example a GPU (Global Processing Unit) and/or a TPU (Tensor Processing Unit) and/or an NPU (Neural Processing Unit), with the use of which as part of machine learning the process can be carried out more quickly. The memory unit 403 can be realized as non-permanent working memory (Random Access Memory, abbreviated to RAM) or as permanent mass memory (hard disk, USB stick, SD card, Solid State Disk). The interface 401 can in particular comprise a number of sub interfaces, which carry out different steps of the respective method. The processor unit 402 can in particular comprise a number of sub processor units, which carry out different steps of the respective method.

(39) Shown schematically in FIG. 5 is the training system 500. The training system 500 shown is embodied to carry out an disclosed method for provision of a result dataset. The training system 500 comprises a training interface 501, a training processing unit 502 and a training memory unit 503.

(40) The training system 500 can in particular involve a computer, a microcontroller or an integrated circuit. As an alternative the training system 500 can involve a real or virtual group of computers (a real group is a cluster and a virtual group is a cloud). The training system 500 can also be embodied as a virtual system, which is executed on a real computer or on a real or virtual group of computers (the technical term is virtualization).

(41) The training interface 401 can involve a hardware interface or software interface (for example PCI bus, USB or Firewire). The training processor unit 402 can have hardware elements or software elements, for example a microprocessor or what is known as an FPGA (acronym for Field Programmable Gate Array). Moreover the training processor unit 402 can comprise components that are specialized and/or optimized for tasks and/or applications for machine learning, such as for example a GPU (Global Processing Unit) and/or a TPU (Tensor Processing Unit) and/or an NPU (Neural Processing Unit), with the use of which as part of machine learning the process can be carried out more quickly. The training memory unit 403 can be realized as non-permanent working memory (Random Access Memory, abbreviated to RAM) or as permanent mass memory (hard disk, USB stick, SD card, Solid State Disk). The training interface 501 can in particular comprise a number of sub interfaces, which carry out different steps of the respective method. The training processor unit 402 can in particular comprise a number of sub processor units, which carry out different steps of the respective method.

(42) Shown schematically in FIG. 6 is the magnetic resonance apparatus 601. The magnetic resonance apparatus 601 comprises a scanner unit 602 formed by a magnet unit. Moreover the magnetic resonance apparatus 601 has a patient accommodation area 603 for accommodating a patient 604. In the present exemplary aspect the patient accommodation area 603 is embodied in a cylindrical shape and is surrounded in a circumferential direction in a cylindrical shape by the scanner unit 602, in particular by the magnet unit. Basically however an aspect of the patient accommodation area 603 differing therefrom is always conceivable. The patient 604 can be pushed and/or moved by means of a patient accommodation apparatus 605 of the magnetic resonance apparatus 601 into the patient accommodation area 603. For this purpose the patient accommodation apparatus 605 has a patient couch 606 embodied movably within the patient accommodation area 603. In particular the patient couch 606 is movably supported here in the direction of a longitudinal extent of the patient accommodation area 603 and/or in the z-direction.

(43) The scanner unit 602, in particular the magnet unit, comprises a superconducting basic magnet 607 for creating a strong and in particular constant basic magnetic field 608. Furthermore the scanner unit 602, in particular the magnet unit, has a gradient coil unit 609 for creation of magnetic field gradients, which are used for spatial encoding during imaging. The gradient coil unit 609 is controlled by means of a gradient control unit 610 of the magnetic resonance apparatus 601. The scanner unit 602, in particular the magnet unit, furthermore comprises a radio-frequency antenna unit 611 for exciting a polarization, which is set up in the basic magnetic field 608 created by the basic magnet 607. The radio-frequency antenna unit 611 is controlled by a radio-frequency antenna control unit 612 of magnetic resonance apparatus 601 and radiates radio-frequency magnetic resonance sequences into the patient accommodation area 603 of the magnetic resonance apparatus 601.

(44) The magnetic resonance apparatus 601 furthermore comprises a local radio-frequency coil 600 for receiving a magnetic resonance signal. To this end the local radio-frequency coil 600 is arranged around a region of the patient 604 to be examined. Preferably the local radio-frequency coils 600 are specifically designed for one examination area of the patient, such as for example radio-frequency head coil to acquire magnetic resonance signals during an examination of the head or a radio-frequency knee coil to acquire magnetic resonance signals during an examination of the knee etc.

(45) For control of the basic magnet 607, of the gradient control unit 610 and for control of the radio-frequency antenna control unit 612 the magnetic resonance apparatus 601 has a system control unit 613. The system control unit 613 centrally controls the magnetic resonance apparatus 601, such as for example the carrying out of a predetermined imaging gradient echo sequence. Moreover the system control unit 613 comprises an evaluation unit not shown in any greater detail for an evaluation of medical image data, which is acquired during the magnetic resonance examination.

(46) The magnetic resonance apparatus 601 furthermore comprises the provision system 400, which is connected to the system control unit 613. As an alternative to the exemplary aspect shown it is also possible for the provision system 400 to be embodied as part of the system control unit 613.

(47) The magnetic resonance apparatus 601 furthermore comprises a user interface 614, which is connected to the system control unit 613. Control information such as for example imaging parameters, as well as reconstructed magnetic resonance images, can be displayed on a display unit 615, for example on at least one monitor, of the user interface 614 for medical operating personnel. The user interface 614 furthermore has an input unit 616, by means of which information and/or parameters can be entered during a measurement process by the medical operating personnel.

(48) The magnetic resonance apparatus 601 shown can of course comprise further components that magnetic resonance apparatuses 601 usually have. The general way in which a magnetic resonance apparatus 601 functions is moreover known to the person skilled in the art, so that a more detailed description of the further components will be dispensed with here.

(49) Although the disclosure has been illustrated and described in detail by the preferred exemplary aspects, the disclosure is not restricted by the disclosed examples and other variations can be derived herefrom by the person skilled in the art, without departing from the scope of protection of the disclosure.