Magnetic resonance imaging system generating anti-noise

Abstract

Disclosed herein is a magnetic resonance imaging system 100 controlled by a processor 130. Execution of the machine executable instructions causes the processor to receive a selection input of a gradient coil pulse command, provide the selected command and at least one value associated with a further parameter to a trained machine learning system 122, and receive information from the machine learning system information regarding an anti-noise to be generated by an acoustic transducer 124, 129 to compensate for a noise experienced at an ear of a subject 118 in the magnetic resonance imaging system. The machine executable instructions further cause the processor to control the magnetic resonance imaging system with a set of pulse sequence commands and gradient coil pulse commands to acquire imaging k-space data, and to operate the acoustic transducer in synchronization therewith to generate the anti-noise using the information output by the trained machine learning system.

Description

[Technical field]

[0001] The present invention relates to magnetic resonance imaging, and in particular to a magnetic resonance imaging system that generates anti-noise by means of an acoustic transducer to compensate for noise resulting from the operation of the magnetic resonance imaging system and experienced at the ears of a subject within the magnetic resonance imaging system. [Background technology]

[0002] A large static magnetic field is used by magnetic resonance imaging (MRI) scanners to align the nuclear spins of atoms as part of a procedure to produce images within a subject's body. This large static magnetic field is called the B0 field or main magnetic field. In spatial encoding, a magnetic field gradient coil system is used to superimpose time-dependent gradient magnetic fields onto the B0 field. The gradient magnetic fields are generated by supplying electric currents to the magnetic field gradient coils. The changes in electric current typically result in a loud, repeating audible noise during a magnetic resonance imaging examination.

[0003] From DE 10 2013 219 309 A1 a method for suppressing such noise in a magnetic resonance imaging system is known. An anti-noise signal is played for a subject in the bore of the magnetic resonance imaging system and controlled, for example, using headphones worn by the patient. That is to say that a pre-determined anti-noise pattern is used, which is stored in a memory, for example in audio format. The anti-noise is intended to represent an inverted real noise. The pattern is obtained as a result of a pre-measurement of the real noise in the bore depending on the sequence used to supply current to the magnetic field gradient coils of the magnetic resonance imaging system, and the closest suitable pattern is later selected in conjunction with the selection of such a sequence. Summary of the Invention

[0004] The invention provides a method for generating anti-noise in a first magnetic resonance imaging system, a magnetic resonance imaging system and a computer program product in the independent claims. Embodiments are provided in the dependent claims.

[0005] During a magnetic resonance imaging scan, loud noise generated by a gradient coil system may startle a subject and cause the subject to move. This results in artifacts and blurring in the resulting magnetic resonance image. Moreover, the subject feels uncomfortable due to this loud noise and earplugs or mufflers worn during the scan. To help reduce or eliminate this, an embodiment provides a system that generates anti-noise with an acoustic transducer to compensate for the noise that originates from the operation of the magnetic resonance imaging system and is experienced at the ears of a subject in the magnetic resonance imaging system. However, it is not an easy task. The noise cannot be measured on the fly (when done outside the field of the magnetic resonance imaging system), then directly inverted, and the inverted noise can be used as anti-noise for compensation purposes. The noise pattern of the magnetic resonance imaging system is too complex, and the noise is changing too quickly for such a procedure. It is difficult to sense the noise at run-time, process and invert it, and play it back in sync with the real noise. The actual noise generated and experienced at the ear of a subject in a magnetic resonance imaging system also depends on many factors beyond gradients, such as system hardware conditions and gradient mountings, and cannot be easily predicted by simple means in a reliable manner.

[0006] The embodiment provides a reliable prediction by using two approaches: First, it does not only rely on the sole dependence of the noise on the gradient coil pulse command, but rather references both the gradient coil pulse command and at least one further parameter describing the imaging of the subject; Second, a trained machine learning system is used to predict the noise, i.e., to provide information on which anti-noise should be used.

[0007] In one aspect, the present invention provides a method for generating anti-noise in a first magnetic resonance imaging system including a bore for receiving a subject to be imaged, the first magnetic resonance imaging system being configured to acquire imaging k-space data from an imaging zone defined in the bore, the first magnetic resonance imaging system including a magnetic field gradient coil system configured to generate a magnetic gradient field within the imaging zone, and a memory including pulse sequence commands configured to control the first magnetic resonance imaging system to acquire imaging k-space data according to a magnetic resonance imaging protocol, the memory further including gradient coil pulse commands configured to control the magnetic field gradient coil system during acquisition of the imaging k-space data. The method includes receiving a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands, and providing the selected gradient coil pulse commands and at least one value associated with each at least one further parameter describing imaging of the subject to a machine learning system to be trained. In response to the providing step, the method includes receiving information from the machine learning system regarding an anti-noise to be generated by the acoustic transducer to compensate for noise experienced at the ear of the subject and resulting from operation of the magnetic resonance imaging system with the first set of gradient coil pulse commands under at least one value constraint. The method further includes controlling the first magnetic resonance imaging system with the pulse sequence commands and the first set of gradient coil pulse commands to acquire imaging k-space data, and operating the acoustic transducer to generate the anti-noise using information output by the trained machine learning system synchronized with control of the magnetic field gradient coil system with the selected gradient coil pulse commands.

[0008] According to one embodiment, the at least one further parameter comprises one or more parameters defining the magnetic field strength B0 of the static main magnetic field of the first magnetic resonance imaging system. The noise is caused when the control of the magnetic field gradient coil system is performed, but the intensity, pitch and pattern of the noise have a more or less strong dependence on the basic static magnetic field B0. The relevant parameter values ​​are generally stored in a memory as information about the magnetic resonance imaging system itself. The machine learning system to be trained has all been trained with magnetic resonance imaging systems having different field strengths B0, or at least only partially different field strengths B0, so that the respective information is used to help predict the noise more accurately.

[0009] According to one embodiment, the at least one further parameter includes one or more parameters that define the general configuration of the magnetic field gradient. This relates to the maximum amplitude (voltage), the maximum slew rate (how fast the gradient is turned on and off, i.e. the maximum gradient strength / rise time measured in millitesla / second for example), and Grms, i.e. the root mean square of the gradient current sent to the magnetic field gradient coil system measured in amperes. These values ​​are not directly related to the gradient coil pulse command, and therefore it is useful to refer to them so that the actual noise can be better predicted. The relevant parameter values ​​are generally stored in memory as information about the magnetic resonance imaging system itself.

[0010] According to another embodiment, the at least one further parameter comprises one or more parameters indicative of the scan commands in the memory used to acquire the k-space data. These scan commands relate to the echo time TE and the repetition time TR, both measured for example in terms of a 64-pixel data package. The scan commands further relate to the resolution, for example indicated by Xres and Yres, the image quality, for example indicated by the readout bandwidth RBW, the slice thickness, the NSA (number of average samples) and the FOV (field of view). All these parameters are well known as indicators of the characteristics of the operation of a magnetic resonance imaging system, each of them having an influence on the noise generated. The relevant parameter values ​​are stored in the memory, for example as a result of a respective input of the operator for a given scan. Such an input generally consists in selecting from a number of options stored in the memory.

[0011] According to another embodiment, the at least one further parameter includes one or more parameters indicative of the subject's relative orientation with respect to the bore. Such indication is made by reference to the anatomical part of the subject (patient) in the bore of the magnetic resonance imaging system. For example, there are a number of options, e.g. "head", "neck", "chest", "abdomen", "pelvis", "feet", "extremities", etc., that are typically stored in a memory, and the parameter is indicated by a numerical value (e.g. 1-7) for the option selected by input of the operator (or a system that automatically detects the subject's anatomical part). The subject's relative orientation with respect to the bore is also indicated by a measured parameter, such as the depth of how deeply the subject is inserted. This embodiment is based on the finding that the noise generated in the bore depends not only on the magnetic resonance imaging system itself, but also on how the subject is positioned in the bore.

[0012] According to another embodiment, the at least one further parameter comprises one or more parameters indicating or defining the subject's physical shape. This is an indication of the subject's gender, height and / or weight. The parameter relating to gender is indicated by a numerical value (e.g. 1 or 2, or 1, 2, 3 or 1, 2, 3, 4) for an option from options in the memory selected by input of the operator (or of the system for automatically detecting the subject). The parameters relating to the subject's height and / or weight are indicated as integer or real values ​​by input of the operator (or of the system for automatically detecting the subject). This embodiment is based on the finding that the noise generated in the bore depends not only on the magnetic resonance imaging system itself, but also on the subject in comparison to another subject.

[0013] According to another embodiment, the at least one further parameter comprises all of the parameters from the parameters listed above. Thus, all of those parameters that have an impact on noise are taken into account and used for and by the machine learning system to be trained. It should be noted that the machine learning system to be trained is intended for a specific type of magnetic resonance imaging system, and therefore some of those parameters listed above are omitted. For example, if the magnetic field strength B0 of the static main magnetic field of the first magnetic resonance imaging system is always the same in training and in later use, the respective information is not necessarily required. On the other hand, the machine learning system is intended for a specific type of magnetic resonance imaging system that includes different bore lengths or heights, and therefore these parameters are also used during training and later used in the trained machine learning system.

[0014] According to another embodiment, at least one further parameter is received via an input device of the magnetic resonance imaging system. Thus, noise is correctly treated as being dependent on, for example, user input. Alternatively, the input may be in an automated manner, receiving input from, for example, a medical information system or another system coupled to the MRI system. Another example of such a system is a medical practice management software (PMS), a category of health management software that handles the day-to-day operations of a medical practice.

[0015] According to another embodiment, the acoustic transducer comprises any of the following: a loudspeaker inside the bore of the first magnetic resonance imaging system, an earphone worn by the subject, a headset worn by the subject, and headphones worn by the subject, each of these options making it possible to provide good positioning of the anti-noise to the ear of the subject.

[0016] In another embodiment, the method includes providing a trained machine learning system by receiving a training data set including a combination of a training set of gradient coil pulse commands and training values ​​associated with parameters describing the imaging of a training subject with noise under the constraints of the training values ​​experienced at the ears of the training subject and generated over time when operating a training magnetic resonance imaging system with the training set of gradient coil pulse commands, and training a machine learning model using the training set of gradient coil pulse commands, the training values, and the noise as input, the training resulting in a trained machine learning system. Thus, the method does not rely on a pre-determined trained machine learning system but includes the training itself.

[0017] According to another embodiment, the training values ​​relate to one or more of those parameters discussed above, for which other values ​​are subsequently used as inputs to the machine learning system being trained. Using the same parameters in training as those of the later input intended to obtain the output is generally advantageous and leads to higher accuracy.

[0018] According to another embodiment, the training magnetic resonance imaging system is a simulated magnetic resonance imaging system and the noise is simulated noise of the simulated magnetic resonance imaging system, where the use of a simulation tool can be proven to be economical and at the same time sufficiently accurate.

[0019] According to another embodiment, the training magnetic resonance system corresponds to the first magnetic resonance system, and thus the machine learning system to be trained is trained during normal use of the first magnetic resonance system, which is economical and results in an increasingly sophisticated and accurate machine learning system.

[0020] According to another embodiment, the training magnetic resonance imaging system is a second magnetic resonance imaging system and the noise is the measured noise of the second magnetic resonance imaging system. This separation of training and use of the machine learning system helps eliminate artifacts, is effective in practical terms, and is primarily the most economical.

[0021] In another aspect, the present invention provides a magnetic resonance imaging system configured to acquire imaging k-space data from an imaging zone within a bore of the magnetic resonance imaging system for receiving a subject to be imaged, the magnetic resonance imaging system including a magnetic field gradient coil system configured to generate a magnetic gradient field within the imaging zone. The magnetic resonance imaging system includes a memory including pulse sequence commands configured to control a first magnetic resonance imaging system to acquire imaging k-space data according to a magnetic resonance imaging protocol. The memory further includes gradient coil pulse commands configured to control the magnetic field gradient coil system during acquisition of the imaging k-space data. The memory still further includes machine executable instructions. The magnetic resonance imaging system further includes at least one acoustic transducer configured to output anti-noise and a processor configured to control the magnetic resonance imaging system, execution of the machine executable instructions causing the processor to receive a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands. Execution of the machine executable instructions further causes the processor to provide the selected gradient coil pulse commands and at least one value associated with each at least one further parameter describing the imaging of the subject to a trained machine learning system external to the processor. Execution of the machine-executable instructions further causes the processor to receive (in response to providing) from the machine learning system information regarding anti-noise to be generated by the acoustic transducer to compensate for noise experienced at the subject's ear and resulting from operation of the magnetic resonance imaging system with the first set of gradient coil pulse commands under at least one value constraint. Execution of the machine-executable instructions further causes the processor to control the first magnetic resonance imaging system with the pulse sequence commands and the first set of gradient coil pulse commands to acquire imaging k-space data.Execution of the machine-executable instructions further causes the processor to operate the acoustic transducer to generate anti-noise using information output by the trained machine learning system synchronized with control of the magnetic field gradient coil system with selected gradient coil pulse commands.

[0022] According to another embodiment, at least one of the at least one values ​​refers to preparatory scan commands, which have a significant impact on noise.

[0023] According to another embodiment, execution of the machine-executable instructions causes the processor to reconstruct the magnetic resonance imaging data from the imaging k-space data, thus this task remaining internal.

[0024] According to another embodiment, the trained machine learning system is realized by a neural network that is part of a magnetic resonance imaging system, where the trained machine learning system is most readily available and at hand.

[0025] According to another embodiment, the magnetic resonance imaging system is configured to be coupled to a neural network implementing a trained machine learning system external to the magnetic resonance imaging system, where the trained machine learning system can then be used by multiple magnetic resonance imaging systems and more easily further trained using an external training data set.

[0026] In another aspect, the present invention provides a medical system, the medical system comprising: a memory storing machine-executable instructions; and a processor configured to control the medical system, wherein execution of the machine-executable instructions causes the processor to control the medical system to provide a trained machine learning module trained to output to the magnetic resonance imaging system information regarding anti-noise to be generated by an acoustic transducer of the magnetic resonance imaging system in response to input of selected gradient coil pulse commands for the magnetic resonance imaging system, and at least one value associated with a respective at least one further parameter describing imaging of the training subject with noise experienced at the ear of the training subject in the magnetic resonance imaging system; Preparing a machine learning model, providing training data including a combination of a training set of gradient coil pulse commands and training values ​​associated with parameters describing the imaging of the training subject with noise under constraints of the training values ​​as experienced at the training subject's ears or otherwise by the training subject within a magnetic resonance imaging system and generated over time when operating the training magnetic resonance imaging system with the training set of gradient coil pulse commands; training a machine learning model using a training set of gradient coil pulse commands, the training values ​​as inputs, and noise as output, where the training results in a trained machine learning system; Includes.

[0027] The training subject may be a live subject or may be a dummy device, such as a water container.

[0028] Such a medical system suitably provides a trained machine learning model for use in a magnetic resonance imaging system according to another aspect. The medical system is also here a magnetic resonance imaging system, and the transducer here has the function of a microphone. In particular, the medical system itself provides the training data used in training the machine learning model.

[0029] Using the trained machine learning system includes providing the selected gradient coil pulse commands and at least one value associated with each at least one further parameter describing imaging of the subject to the trained machine learning system, and in response to the providing, receiving from the machine learning system information regarding anti-noise to be generated by the acoustic transducer to compensate for noise experienced at the subject's ear and resulting from operation of the magnetic resonance imaging system using the first set of gradient coil pulse commands, subject to the constraint of the at least one value.

[0030] In another aspect, a method of obtaining a trained machine learning model for use in a magnetic resonance imaging system is provided, the method comprising the steps of: providing a machine learning model; providing training data including a combination of a training set of gradient coil pulse commands and training values ​​associated with parameters describing the imaging of the training subject with noise under constraints of the training values ​​as experienced at the training subject's ears or otherwise by the training subject within a magnetic resonance imaging system and generated over time when operating the training magnetic resonance imaging system with the training set of gradient coil pulse commands; training a machine learning model using a training set of gradient coil pulse commands, the training values ​​as input, and noise as output, where the training results in a trained machine learning system; has.

[0031] In yet another aspect, the present invention is provided as a computer program comprising machine executable instructions configured to perform a method for obtaining a trained machine learning model.

[0032] In another aspect, the present invention provides a computer program comprising machine executable instructions configured to control a magnetic resonance imaging system to acquire imaging k-space data from an imaging zone, the magnetic resonance imaging system comprising a magnetic field gradient coil system configured to generate a magnetic gradient field in the imaging zone, and a memory comprising pulse sequence commands configured to control a first magnetic resonance imaging system to acquire imaging k-space data according to a magnetic resonance imaging protocol, the memory further comprising gradient coil pulse commands configured to control the magnetic field gradient coil system during acquisition of the imaging k-space data. Execution of the machine executable instructions causes the processor to receive a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands. Execution of the machine executable instructions further causes the processor to provide the selected gradient coil pulse commands and at least one value associated with each at least one further parameter describing imaging of the subject to a machine learning system external to the processor. Execution of the machine-executable instructions further causes the processor to receive (in response to providing) information from the machine learning system regarding an anti-noise to be generated by the acoustic transducer to compensate for noise experienced at the subject's ear and resulting from operation of the magnetic resonance imaging system with the first set of gradient coil pulse commands under at least one value constraint. Execution of the machine-executable instructions further causes the processor to control the first magnetic resonance imaging system with the pulse sequence commands and the first set of gradient coil pulse commands to acquire imaging k-space data. Execution of the machine-executable instructions further causes the processor to operate the acoustic transducer to generate the anti-noise using information output by the trained machine learning system synchronized with control of the magnetic field gradient coil system with the selected gradient coil pulse commands.

[0033] It should be understood that one or more of the above-described embodiments of the present invention may be combined, unless the combined embodiments are mutually exclusive.

[0034] As will be appreciated by those skilled in the art, aspects of the invention may be embodied as an apparatus, a method, a computer program, or a computer program product. Accordingly, aspects of the invention may take the form of an entirely hardware embodiment, an entirely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, all generally referred to herein as a "circuit," "module," or "system." Additionally, aspects of the invention may take the form of a computer program product embodied in one or more computer-readable medium(s) having computer-executable code embodied therein. A computer program comprises computer-executable code or "program instructions."

[0035] Any combination of one or more computer readable media may be utilized. A computer readable medium may be a computer readable signal medium or a computer readable storage medium. As used herein, a "computer readable storage medium" encompasses any tangible storage medium that stores instructions executable by a processor of a computing device. A computer readable storage medium may be referred to as a computer readable non-transitory storage medium. A computer readable storage medium may also be referred to as a tangible computer readable medium. In some embodiments, a computer readable storage medium may also store data that may be accessed by a processor of a computing device. Examples of computer readable storage media include, but are not limited to, floppy disks, magnetic hard disk drives, solid state hard disks, flash memory, USB memory, random access memory (RAM), read only memory (ROM), optical disks, magneto-optical disks, and processor register files. Examples of optical disks include compact disks (CDs) and digital versatile disks (DVDs), such as CD-ROM, CD-RW, CD-R, DVD-ROM, DVD-RW, or DVD-R disks. The term computer-readable storage medium also refers to various types of recording media that can be accessed by a computer device over a network or communication link. For example, data may be retrieved over a modem, over the Internet, or over a local area network. Computer executable code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, fiber optic cable, RF, or the like, or any suitable combination of the foregoing.

[0036] A computer-readable signal medium includes a propagated data signal in which computer-executable code is embodied, for example, in baseband or as part of a carrier wave. Such a propagated signal may take any of a variety of forms, including, but not limited to, electrical, magnetic, optical, or any suitable combination thereof. A computer-readable signal medium is not a computer-readable storage medium, but is any computer-readable medium that communicates, propagates, or transports a program for use by or in connection with an instruction execution system, apparatus, or device.

[0037] "Computer memory" or "memory" is one example of a computer-readable storage medium. Computer memory is any memory that is directly accessible by a processor. "Computer storage" or "storage" is a further example of a computer-readable storage medium. Computer storage is any non-volatile computer-readable storage medium. In some embodiments, computer storage is also computer memory, or vice versa.

[0038] As used herein, a "processor" encompasses an electronic component capable of executing a program or machine-executable instructions or computer-executable code. References to a computing device including a "processor" should be interpreted as including more than one processor or processing core, as the case may be. A processor is, for example, a multi-core processor. A processor also refers to a collection of processors within a single computer system or distributed among multiple computer systems. The term computing device should also be interpreted as referring to a collection or network of computing devices, each of which includes one or more processors, as the case may be. Computer-executable code is executed by multiple processors, either within the same computing device or even distributed across multiple computing devices.

[0039] Computer executable code includes machine executable instructions or programs that cause a processor to perform aspects of the present invention. Computer executable code for performing operations related to aspects of the present invention may be written in any combination of one or more programming languages, including object oriented programming languages ​​such as Java, Smalltalk, C++, and the like, and conventional procedural programming languages ​​such as the "C" programming language or similar programming languages, and compiled into machine executable instructions. In some cases, the computer executable code is in the form of a high level language or in a pre-compiled form and is used in conjunction with an interpreter that generates the machine executable instructions on the fly.

[0040] The computer executable code may run entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer through any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or the connection may be made to an external computer (e.g., through the Internet using an Internet Service Provider).

[0041] Generally, the program instructions are executed on one processor or on several processors. In case of multiple processors, they may be distributed across several different entities such as clients, servers, etc. Each processor executes a part of the instructions intended for that entity. Thus, when referring to a system or process including multiple entities, it is understood that the computer program or program instructions are adapted to be executed by a processor associated with or associated with the respective entity.

[0042] Aspects of the present invention are described with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It will be understood that each block or part of the blocks of the flowcharts, diagrams, and / or block diagrams, where applicable, are implemented by computer program instructions in the form of computer executable code. It will be further understood that combinations of blocks in different flowcharts, diagrams, and / or block diagrams may be combined, if not mutually exclusive. These computer program instructions are provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the instructions, executed via the processor of the computer or other programmable data processing apparatus, create means for performing the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams to create a machine.

[0043] These computer program instructions may also be stored on a computer-readable medium that directs a computer, other programmable data processing apparatus, or other device to function in a particular manner, such that the instructions stored on the computer-readable medium produce an article of manufacture including instructions that implement the functions / acts specified in one or more blocks of the flowcharts and / or block diagrams.

[0044] Computer program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device to create a computer-implemented process, a sequence of operational steps to be executed on the computer, other programmable apparatus, or other device, such that the instructions executing on the computer or other programmable apparatus provide a process for implementing the functions / operations specified in one or more blocks of the flowcharts and / or block diagrams.

[0045] A "user interface" as used herein is an interface that allows a user or operator to interact with a computer or computer system. A "user interface" is also referred to as a "human interface device." A user interface provides information or data to an operator and / or receives information or data from an operator. A user interface can allow a computer to receive input from an operator and provide output from the computer to a user. In other words, a user interface allows an operator to control and manipulate a computer, and the interface allows a computer to show the effect of the operator's control or manipulation. Displaying data or information on a display or graphical user interface is an example of providing information to an operator. Receiving data via a keyboard, mouse, trackball, touchpad, pointing stick, graphics tablet, joystick, gamepad, webcam, headset, pedals, wired gloves, remote control, and accelerometer are all examples of user interface components that allow receiving information or data from an operator.

[0046] As used herein, a "hardware interface" encompasses an interface that allows a processor of a computing system to interact with and / or control external computing devices and / or devices. A hardware interface allows a processor to send control signals or instructions to external computing devices and / or devices. A hardware interface also allows a processor to exchange data with external computing devices and / or devices. Examples of hardware interfaces include, but are not limited to, a universal serial bus, an IEEE 1394 port, a parallel port, an IEEE 1284 port, a serial port, an RS-232 port, an IEEE-488 port, a Bluetooth connection, a wireless local area network connection, a TCP / IP connection, an Ethernet connection, a control voltage interface, a MIDI interface, an analog input interface, and a digital input interface.

[0047] As used herein, a "medical system" encompasses any system including a memory storing machine-executable instructions and a computing system, e.g., a processor configured to execute the machine-executable instructions, where execution of the machine-executable instructions causes the computing system to provide or process a medical image. The medical system is configured to use the trained machine learning module to train the machine learning module to provide information regarding the anti-noise to be generated. The medical system is further configured to generate a medical image using the medical imaging data and / or to acquire medical imaging data for generating a medical image.

[0048] K-space data is defined herein as being the recorded measurements of radio frequency signals emitted by atomic spins using the antenna of a magnetic resonance device during a magnetic resonance imaging scan. Magnetic resonance data is an example of medical image data. A magnetic resonance imaging (MRI) image or MR image is defined herein as being a reconstructed two- or three-dimensional visualization of the anatomical data contained within the k-space data. This visualization is performed using a computer.

[0049] Preferred embodiments of the present invention will now be described, by way of example only, with reference to the accompanying drawings, in which: [Brief description of the drawings]

[0050] [Figure 1] FIG. 1 illustrates an example of a magnetic resonance imaging system. [Diagram 2] FIG. 1 illustrates another example of a magnetic resonance imaging system. [Diagram 3] FIG. 1 is a diagram illustrating the operating principle of noise cancellation. [Figure 4] FIG. 1 is a schematic diagram showing an example of generating anti-noise. [Diagram 5] FIG. 1 illustrates another example of a magnetic resonance imaging system. [Figure 6] FIG. 1 illustrates a machine learning system used during training. [Figure 7] FIG. 7 illustrates the machine learning system of FIG. 6 in use after training. [Figure 8a] 1 is a flow diagram illustrating an example of a method for generating anti-noise. [Figure 8b] 1 is a flow diagram illustrating only a method for obtaining a trained machine learning system. [Figure 8c] FIG. 1 is a flow diagram illustrating only a method for generating anti-noise, where a trained machine learning system is already available. [Figure 9a] FIG. 1 illustrates an example of a headset. [Figure 9b] FIG. 13 illustrates another example of a headset. [Figure 9c] FIG. 1 is a top cross-sectional view showing the arrangement of multiple loudspeakers within the bore of a magnetic resonance imaging system. [Figure 9d] FIG. 9b is a side view showing the arrangement of FIG. 9c. [Figure 10a] FIG. 13 is a diagram showing a noise pattern of real noise. [Figure 10b] FIG. 13 is a diagram showing a noise pattern of synthesized noise. [Figure 10c] FIG. 13 is a diagram showing the frequency structure of real noise. [Figure 10d] FIG. 13 is a diagram showing the frequency structure of synthesized noise. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0051] Elements that are numbered the same in the figures are equivalent elements or perform the same function. An element that is previously discussed is not necessarily discussed in a later figure if the function is equivalent.

[0052] 1 and 2 both show an example of a magnetic resonance imaging system 100. The magnetic resonance imaging system includes a magnet 102. The magnet 102 is, for example, a superconducting magnet. Alternatively, the magnet 102 is a resistive type magnet.

[0053] Different types of magnets can be used, for example both split cylindrical magnets and so-called open magnets. Split cylindrical magnets are similar to standard cylindrical magnets, except that the cryostat is split into two sections to allow access to the magnet's isoplane, such magnets are used, for example, in conjunction with charged particle beam therapy. Open magnets have two magnet sections, one above the other, with a space between them large enough to accommodate the subject, and the arrangement of the two section areas is similar to that of a Helmholtz coil. Open magnets are popular because they are less confining to the subject. Inside the cryostat of the cylindrical magnet is a collection of superconducting coils. Within the bore 106 of the cylindrical magnet 102 is the imaging zone 108, where the magnetic field is strong and uniform enough to perform magnetic resonance imaging. A region of interest 109 is shown within the imaging zone 108. The subject 118 is shown as being supported by a subject support 120 such that at least a portion of the subject 118 (i.e., the anatomical tissue of the subject 118 that needs to be imaged) is within the imaging zone 108 and the region of interest 109.

[0054] Also present within the magnet bore 106 are a set of magnetic field gradient coils 110 used for acquiring preliminary magnetic resonance data for spatially encoding magnetic spins within the imaging zone 108 of the magnet 102. The magnetic field gradient coils 110 are connected to a magnetic field gradient coil power supply 112. The magnetic field gradient coils 110 are intended to be representative. Typically, the magnetic field gradient coils 110 include a set of three separate coils for spatially encoding in three orthogonal spatial directions. The magnetic field gradient power supply supplies current to the magnetic field gradient coils 110. The current supplied to the magnetic field gradient coils 110 is controlled, gradient or pulsed as a function of time.

[0055] Adjacent to the imaging zone 108 is a radio frequency coil 114 for receiving radio transmissions from the spins in the imaging zone 108. In some examples, the radio frequency coil is also configured to manipulate the direction of the magnetic spins in the imaging zone 108. The radio frequency antenna includes multiple coil elements. The radio frequency antenna is also called a channel or antenna. The radio frequency coil 114 is connected to a radio frequency receiver or transceiver, not shown here. The radio frequency coil 114 and the radio frequency transceiver may optionally be replaced by separate transmit and receive coils and separate transmitters and receivers. It should be understood that the radio frequency coil 114 is representative. The radio frequency coil 114 may also represent a dedicated transmit antenna and a dedicated receive antenna. Similarly, separate transmitters and receivers may also be present. The radio frequency coil 114 may also have multiple receive / transmit elements, and the radio frequency transceiver may have multiple receive / transmit channels. For example, if a parallel imaging technique such as SENSE is performed, the radio frequency coil 114 has multiple coil elements.

[0056] The magnetic resonance imaging system 100 is illustratively shown as including a machine learning system 122. However, alternatively, the machine learning system 122 is external to the system 100. The machine learning system 122 is used to provide anti-noise. The machine learning system 122 is shown as being coupled to a set of acoustic transducers, e.g., loudspeakers 124 and / or headphones 129. Both the acoustic transducers 124 and the headphones 125 are used to provide anti-noise to noises generated during operation of the magnetic resonance imaging system 100. In all embodiments, not both the loudspeakers 124 and the headphones 129 are required, one of them being sufficient. Here, only a single loudspeaker 124 is shown, but there may also be four loudspeakers 124 in the bore, e.g., two of the pair at one (horizontal) end and two of the pair at the other (horizontal) end, and the loudspeakers 124 are opposed in pairs, one at the top and one at the bottom. There are other arrangements with four loudspeakers, and alternatively there may be two, three, five or more loudspeakers. The acoustic transducer 124 and the headphones 129 are constructed using magnetic resonance imaging compatible technology. The machine learning system 122 is shown to include an autoencoder unit 123, a recurrent network unit 125, and a unit 127 responsible for interaction and data exchange between the autoencoder unit 123 and the recurrent network unit 125, and with external entities.

[0057] The gradient controller 112 is shown as being connected to a hardware interface 128 of the computer system 126. Optionally (as shown), but not necessarily in any case, the machine learning system 122 is similarly connected to the hardware interface 128. The computer system further includes a processor 130 in communication with the hardware interface 128, a clock 131, a memory 134, and a user interface 132. The memory 134 is any combination of memory accessible to the processor 130. This includes main memory, cache memory, and even non-volatile memory, such as flash RAM, hard drives, or other storage devices. In some examples, the memory 130 is a non-transitory computer-readable medium.

[0058] The memory 134 is shown as including machine executable instructions 140. The machine executable instructions 140 provide a means for the processor 130 to control the magnetic resonance imaging system 100. The machine executable instructions 140 also enable the processor 130 to perform various data analysis and image reconstruction tasks. The machine executable instructions 140 also enable the processor 130 to control the user interface 132 to provide inputs to be made, for example, by requesting the input (by a mask layer with a display user interface or by audio commands in an audio user interface), i.e., by allowing selection from pre-determined units or numerical input. The machine executable instructions 140 then further enable the processor 130 to receive these inputs and have the input data stored in the memory 134.

[0059] The memory 134 is further shown as containing a number of preparatory scan commands 142. The memory 134 is then shown as containing a selected preparatory scan command 144.

[0060] The memory 134 is shown as optionally including configuration data 146. The configuration data 146 is data acquired or derived during execution of the preparatory scan commands 144. In some examples, this is used to configure pulse sequence commands 148. The pulse sequence commands 148 are shown as also being stored in the memory 134. The pulse sequence commands 148 are commands or data that are converted into such commands that are used to control the magnetic resonance imaging system to acquire k-space data. The memory 134 is shown as including imaging k-space data 150 acquired by controlling the magnetic resonance imaging system with the pulse sequence commands 148. The memory 134 is further shown as including magnetic resonance imaging data 152 reconstructed from the imaging k-space data 150. The imaging k-space data 150 is acquired, for example, for the region of interest 109.

[0061] The pulse sequence commands 148 include gradient coil pulse commands 154. The gradient coil pulse commands 154 are commands or data that are converted into such commands that control the controlled magnetic field gradient coil power supplies 112 to supply current to the magnetic field gradient coils 110. When this occurs, it creates a loud audible noise within the bore 106 of the magnet 102. The memory 134 is further shown as including a gradient pulse start time 156 and a predetermined time 158. The gradient pulse start time 156 is the time that the execution of the gradient coil pulse commands 158 begins. This is essentially the time that the loud noise begins to be created within the bore 106 of the magnet 102. Knowing what the gradient pulse start time 156 is, the anti-noise is delivered in a predetermined manner, i.e., relative to the predetermined time 158, to synchronize the generation of the anti-noise to the loud audible noise.

[0062] The machine executable instructions 140 provide a means for the processor to control the acoustic transducer 124 and the headphones 129 via the machine learning system 122. That is, the anti-noise signal is preferably always played from the machine learning system 122. The time to start playing the signal is based on a gradient start time 156 received as the machine executable instructions 140 through the hardware interface 128. The machine learning system 122, in most embodiments, communicates with the memory 134 using the hardware interface 128 to control the acoustic transducer 124 and / or the headphones 129 in playing or by playing the anti-noise signal. However, communication also includes other paths. While FIG. 1 shows that the hardware interface 128 is optionally coupled to the gradient controller 112 and the machine learning system 122, FIG. 2 shows that the hardware interface 128 is additionally optionally coupled to the acoustic transducer 124 and the headphones 129.

[0063] The machine learning system receives as input a selection of a first gradient coil pulse command 160 from the gradient coil pulse commands 154 via the user interface 132. The memory further stores a value 162 that is generally related to a parameter of the magnetic resonance imaging system 100 and is constant in the magnetic resonance imaging system 100. Such a further value relates to the strength of the magnet 102 (e.g., in Tesla). A number of possible selectable values ​​164 are likewise stored in the memory, and a value 166 resulting from a related dedicated input via the user interface 132 is likewise stored in the memory 134. For example, the possible selectable values ​​164 include an indication of an anatomical part of the subject 118 in the imaging zone. The operator can select one of the values ​​164 indicating "head" (e.g., "1"), "neck" ("2"), "chest" ("3"), "abdomen" ("4"), "pelvis" ("5"), "feet" ("6"), and "extremities" ("7") and store the selected value 166 in the memory. Moreover, the memory is further shown to store additional values ​​168 (e.g., numerical values) that are freely entered via the user interface 132 without any significant restriction. Such additional values ​​include those indicative of the age, height, and / or weight of the subject 118.

[0064] FIG. 3 shows a schematic of the working principle of noise cancellation in earplugs 170. Here, the impinging noise, represented by sound wave 172, is compensated by the superposition of the same wave, except that it is inverted to form an inverted wave 174. "Inverted" means that the wave is phase shifted by 180°. Thus, the superposition of wave 172 and inverted wave 174 means that the noise is cancelled at 176. Generally, the noise is recorded and inverted, but in this embodiment, the anti-noise is predicted. Then, instead of inversion, the anti-noise wave is directly generated based on the values ​​160, 162, 164, 166 and 168 stored in memory 130 (i.e., by the "predicted noise"), and the effect of canceling is the same. The timing is of course important. Here, the noise is predicted by referring, among other things, to the selected gradient coil pulse command 160. The related control is synchronized with the generation of the anti-noise using the same clock 131.

[0065] 4 shows a schematic example of how the anti-noise is generated. The magnetic field gradient coil 110 is controlled by selected gradient coil pulse commands 160, the respective information of which is provided as input to a machine learning system 122, here shown as an artificial intelligence AI model. Further values, for example one or more of the values ​​162, 164, 166, 168, are likewise provided to the machine learning system 122. The noise 112 in the bore 106 is compensated by the output of the machine learning system 122, i.e. the anti-noise 174, to provide the cancelled noise 176. The machine learning system 122 needs to be trained before such application.

[0066] FIG. 5 shows an example of how the magnetic resonance imaging system 100 (or another magnetic resonance imaging system) is used to train the machine learning system 122. The magnetic resonance imaging system 100 in the context of FIG. 5 differs from the magnetic resonance imaging system 100 in the context of FIG. 1 in that a microphone 180 is placed in the bore to record the noise, and in that a further microphone is placed at 182 instead of the headphones 129, i.e. at the location where noise cancellation is desired. In all embodiments, both microphones 180 and 182 are not necessary, one of them is sufficient. Thus, the real noise 112 produced in the bore 106 is recorded and provided as training data to the machine learning system 122. The training data further includes parameters that affect the noise, such as those realized by values ​​162, 166, 168, and optionally 164. It should be noted that the values ​​162, 164, 166, 168 are present, selected or entered in the magnetic resonance imaging system 100 without the need for the acoustic transducer 124 or the headphones 129. Basically, if the loudspeaker 124 provides anti-noise in the magnetic resonance imaging system 100, data obtained with the microphone 180 is more appropriately used for training. If the headphones 129 provide anti-noise in the magnetic resonance imaging system 100, data obtained with the microphone 182 is more appropriately used for training. Training is performed with training data including sound (noise) recorded by both microphones 180 and 182 so that it can be decided later whether to use the loudspeaker 124 and the headphones 129 or both.

[0067] FIG. 6 shows an example of a machine learning system 122 with an autoencoder network unit 123 and a regression network unit 125 while the machine learning system 122 is being trained. An acoustic signal (or a noise signal) is recorded at 188 and a short-time Fourier transform is applied to the data at 190 (STFT). The resulting data is input at 192 to the autoencoder unit 123. The autoencoder unit 123 is an unsupervised network learning system that imposes a bottleneck of the network that forces a compressed knowledge representation of the original input in the hidden layer 195 (between the encoder 194 and the decoder 196). This requires that there is some dependency on the input, which is the case for the recorded noise. Now, the bottleneck sends the data at output 186 to the regression network 125 to perform training on the regression network 125. The magnetic field gradient coil 110 is controlled by using selected gradient coil pulse commands 160, and the values ​​of the respective parameters and further parameters are input at 184. These are input to a Recurrent Network unit 125 (this and all the latter are controlled by unit 127), which outputs at 186 a reduced representation for the so-called bottleneck of the hidden layer 195 of the Autoencoder unit 123. The training output from the autoencoder is likewise provided in STFT format at 197.

[0068] It should be noted that instead of using a "real" MRI system with microphones to record the noise, it is also possible to simulate the occurrence of noise for the variables associated with the given gradients and sequence. Later, this simulated noise as the recorded noise at 188 and the associated variables at 184 are provided to the machine learning system 122 for training.

[0069] Fig. 7 shows an example of how such a trained machine learning system 122 is used to generate an anti-noise. Again, the magnetic field gradient coil 110 is controlled by using the selected gradient coil pulse command 160, and values ​​for the respective parameters and further parameters are input at 184. Now, instead of recording the sound or the noise, the recurrent network 125 outputs at 186 a reduced representation of the bottleneck of the audio encoder unit 123, and this time a prediction for the anti-noise is output at 198 and sent to a "noise compensation stage" 199, where it is "synchronized" with the noise. The noise compensation stage 199 represents the acoustic transducer 124 or the headphones 129, and the anti-noise is thus generated to compensate for the external noise. The noise compensation stage 199 additionally (optionally) includes a processor 130 providing a direct output to the acoustic transducer 124 or headphones 129 according to Fig. 2, or optionally includes any means for converting the output of the audio encoder unit 123 at 198 into a controller signal sent to the acoustic transducer 124 or headphones 129. The noise compensation stage 199 also includes using a signal produced by a clock 131. Such a signal of the clock 131 is used both for timing the control of the magnetic field gradient coil 110 by the gradient coil pulse commands and for timing the generation of anti-noise in synchronism with the timing of the control of the magnetic field gradient coil 110.

[0070] FIG. 8a shows a flow diagram of a method of operating the magnetic resonance imaging system 100. The method starts with obtaining a trained machine learning system in steps 200 and 202, and may also optionally start with step 204 if training has already been performed. FIG. 8b shows a flow diagram of a method of obtaining only a trained machine learning system, for example using the magnetic resonance imaging system of FIG. 5. This includes a step 200b of preparing a machine learning system (generally as a model). The latter step may also be included in the method of FIG. 8a. FIG. 8c shows a flow diagram of a method of generating anti-noise, where the machine learning system is already trained, i.e. already prepared as trained, such a method being performed in the magnetic resonance imaging system of FIG. 1 or FIG. 2. (The method of FIG. 8a is performed in a magnetic resonance imaging system having both a loudspeaker as in FIG. 1 or FIG. 2 and a microphone of FIG. 5).

[0071] In step 200, common to both Figures 8a and 8b, a training data set (a combination of selected gradient pulse commands, at least one value for each parameter, and recorded noise) is received by the machine learning system, which is then trained in step 202.

[0072] The method proceeds from the method of FIG. 8a to the method of FIG. 8c in step 204, starting with the method of FIG. 8c, where a selection input regarding a gradient coil pulse command 160 is received, for example, via the user interface 132. Then, in step 206, the selected gradient coil pulse command and at least one further value are provided to the machine learning system. Information regarding anti-noise is received from the machine learning system in response to step 206. Step 210 involves controlling the magnetic resonance imaging system with the selected gradient coil pulse command, and in step 212, the acoustic transducer (124, or headphones 129) is operated using the information received in step 208. Steps 210 and 212 are performed in parallel, i.e. are synchronized by step 214. However, if the timing is anyway very accurate, step 214 may be omitted.

[0073] 9a shows an example of a headset. The headset, designated 300, includes a clip 302 that is placed on the head and two ear pieces (only one is shown) designated 304. The ear pieces 304 generally include a sponge foam for comfortable wearing and an acoustic transducer (not shown) for creating anti-noise.

[0074] Figure 9b shows another example of a headset. This headset 310 differs from headset 300 in that a sponge 316 is placed under the clip 314, the clip is formed as a hollow tubelet, and the anti-noise exits directly at its outlet 318. The transducer is placed within the tubelet or outside the headset.

[0075] In one embodiment, both headsets 300 and 310 are used.

[0076] Further alternatively, the headset includes a closed bowl- or cup-shaped ear cover for at least one ear of the subject (patient), preferably for both ears, in particular without a sponge part. Inside the cover, the anti-noise is generated or transmitted thereto. If the cover is completely closed, such noise is not as loud as without the headset, and therefore the anti-noise may not be as strong. The cover may have other shapes or may be partially open.

[0077] 9c and 9d show another example setup for playing anti-noise within the magnet bore 106. One or more loudspeakers 320a, 320b, 320c, and 320d are, in this embodiment, permanently placed at strategically selected locations inside the magnetic resonance imaging system 100, and the anti-noise is played from the speakers to cancel the actual noise, thereby avoiding the need for headphones for the patient.

[0078] Fig. 10a shows the noise pattern of the real noise and Fig. 10b shows the noise pattern of the respective synthesized (predicted) noise when an embodiment of the method is used. Fig. 10c shows the respective frequency composition of the noise pattern of Fig. 10a and Fig. 10d shows the respective frequency composition of the noise pattern of Fig. 10b. The frequency composition was obtained by fast Fourier transform FFT. From the comparison, it can be seen that this embodiment is very good at predicting the real noise, thereby making it possible to generate an appropriate anti-noise for the comfort of the subject in the bore 106.

[0079] Examples may include one or more of the following features.

[0080] A system for machine learning that is or is to be trained (signaling system 122).

[0081] A system for presenting anti-noise to a subject, for example a loudspeaker, a headphone, a headset, or an earphone with shielding against the magnetic fields in a magnetic resonance imaging system.

[0082] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive, and the invention is not limited to the disclosed embodiments.

[0083] Other variations to the disclosed embodiments can be understood and effected by those skilled in the art, from a study of the drawings, the disclosure, and the appended claims, in practicing the claimed invention. In the claims, the words "comprise, include, have" do not exclude other elements or steps, and singular elements do not exclude a plurality. A single processor or other unit fulfills the functions of several items recited in the claims. The mere fact that certain means are recited in mutually different dependent claims does not indicate that a combination of these means cannot be used to advantage. The computer program may be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state medium, provided together with or as part of other hardware, but may also be distributed in other forms, such as via the Internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be interpreted as limiting the scope. [Explanation of symbols]

[0084] 100 Magnetic Resonance Imaging System 102 Magnet 106 Magnet Bore 108 Imaging Zone 109 Areas of Interest 110 Magnetic field gradient coil 112 Magnetic field gradient coil power supply 114 Radio Frequency Coil 118 Target 120 Subject Support 122 Machine Learning Systems 123 Audio Encoder Unit 124 Acoustic Transducer 125 Regression Network Unit 126 Computer Systems 127 Interaction Units 128 Hardware Interface 129 Headphones 130 processors 131 Clock 132 User Interface 134 Computer Memory 140 Machine Executable Instructions 142 Multiple Prep Scan Commands 144 Selected Preparation Scan Command 146 Configuration Data 148 Pulse Sequence Commands 150 imaging k-space data 152 Magnetic Resonance Imaging Data 154 Gradient coil pulse command 156 Gradient Pulse Start Time 158 Predetermined Time 160 Selected (first) gradient coil pulse command 162 Stored Constant Values 164 selectable values 166 Selected Values 168 Input Values 170 Earplugs 172 External Noise 174 Inversion Waves as Anti-Noise Signals 176 Cancelled Noise Signal 180 Microphone 182 Headphones 184 Input to Recurrent Network Unit 125 186 Output of Regression Network Unit 125 188 Recording of Acoustic Signals 190 Short-time Fourier transform 192 Input to the autoencoder unit 123 194 Encoder 195 Bottleneck 196 Decoder 197 Training Results 198 Autoencoder network unit 123 output 199 Noise Compensation Stage 200 Receive a training dataset for a machine learning system 202 Training a Machine Learning System 204 Receives selection input regarding gradient coil pulse commands 206 Providing the selected gradient coil pulse command and at least one further value to the machine learning system 208 Receive anti-noise information from the machine learning system 210 Controlling a magnetic resonance imaging system using selected gradient coil pulse commands 212 Operate an acoustic transducer using the information received in step 208 214 Synchronize steps 210 and 212 300 Headset 302 clips 304 Earpiece 310 Headset 314 clips 316 Sponge 318 Exit 320a~320d loudspeakers

Claims

1. A method for generating anti-noise in a first magnetic resonance imaging system including a bore for receiving a subject to be imaged, wherein the first magnetic resonance imaging system is configured to acquire imaging k-space data from an imaging zone defined in the bore, the first magnetic resonance imaging system including a magnetic field gradient coil system for generating a magnetic gradient field within the imaging zone and a memory including pulse sequence commands for controlling the first magnetic resonance imaging system to acquire the imaging k-space data according to a magnetic resonance imaging protocol, the memory further including gradient coil pulse commands for controlling the magnetic field gradient coil system during the acquisition of the imaging k-space data, The method comprising: receiving a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands; providing the selected gradient coil pulse commands and at least one value associated with each at least one further parameter describing the imaging of the subject to a trained machine learning system; receiving, in response to the providing step, information regarding anti-noise to be generated by an acoustic transducer to compensate for noise experienced in the ears of the subject and resulting from the operation of the magnetic resonance imaging system using the first set of gradient coil pulse commands, under the at least one value, wherein the machine learning model is trained using a training set of gradient coil pulse commands, training values as inputs, and the noise as outputs; controlling the first magnetic resonance imaging system using the pulse sequence commands and the first set of gradient coil pulse commands to acquire the imaging k-space data; Operating the acoustic transducer to generate anti-noise using the information output by the trained machine learning system synchronized with the control of the magnetic field gradient coil system using the selected gradient coil pulse command A method having **Claim 2** wherein the at least one additional parameter is a parameter defining the magnetic field strength B0 of the static main magnetic field of the first magnetic resonance imaging system, one or more parameters defining the general configuration of the magnetic field gradient, one or more parameters indicating the scan command in the memory used to acquire k-space data, one or more parameters indicating the relative orientation of the subject with respect to the bore one or more parameters defining the body shape of the subject The method according to claim 1, comprising at least one parameter from the group comprising **Claim 3** The method according to claim 2, wherein the at least one additional parameter comprises all of the parameters from the group. **Claim 4** The method according to any one of claims 1 to 3, wherein the method comprises receiving the at least one additional parameter via an input device of the magnetic resonance imaging system. **Claim 5** The method according to any one of claims 1 to 3, wherein the acoustic transducer comprises any one of a loudspeaker inside the bore of the first magnetic resonance imaging system, earphones worn by the subject, a headset worn by the subject, and headphones worn by the subject. **Claim 6** a) Receiving a training data set including a training set of gradient coil pulse commands and a combination with training values related to parameters describing the imaging of the subject under training having noise over time when operating a training magnetic resonance imaging system using the training set of gradient coil pulse commands, the imaging being experienced by the ears of the subject under training; b) Training a machine learning model using the training set of gradient coil pulse commands, the training values as inputs, and the noise as outputs, the training resulting in the trained machine learning system; Further comprising preparing a trained machine learning system having the steps of any one of claims 1 to 3.

7. The parameters related to the training values are A parameter defining the magnetic field strength B0 of the static main magnetic field of the training magnetic resonance imaging system; One or more parameters defining the general configuration of the magnetic field gradient of the training magnetic resonance imaging system; One or more parameters indicating the scan commands in the memory used to acquire the k-space data; One or more parameters indicating the relative orientation of the subject under training with respect to the bore; One or more parameters defining the body shape of the subject under training; The method according to claim 6, comprising at least one parameter from the group including.

8. The training magnetic resonance imaging system is a simulated magnetic resonance imaging system, and the noise is the simulated noise of the simulated magnetic resonance imaging system; The training magnetic resonance system corresponds to the first magnetic resonance system; The training magnetic resonance imaging system is a second magnetic resonance imaging system, and the noise is the measured noise of the second magnetic resonance imaging system The method according to claim 6, wherein the method is any one of .

9. A magnetic resonance imaging system for acquiring imaging k-space data from an imaging zone within a bore of the magnetic resonance imaging system for receiving a subject to be imaged, the magnetic resonance imaging system comprising: A magnetic field gradient coil system for generating a magnetic gradient field within the imaging zone; A memory including pulse sequence commands for controlling the first magnetic resonance imaging system to acquire the imaging k-space data according to a magnetic resonance imaging protocol, the memory further including gradient coil pulse commands for controlling the magnetic field gradient coil system during the acquisition of the imaging k-space data, the memory including machine-executable instructions; At least one acoustic transducer for outputting anti-noise; A processor for controlling the magnetic resonance imaging system, wherein, by executing the machine-executable instructions, the processor: Receiving a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands; Providing the selected gradient coil pulse command and at least one value associated with each at least one further parameter describing the imaging of the subject to a machine learning system to be trained, the machine learning model being trained using a training set of gradient coil pulse commands, training values as inputs, and the noise as outputs; In response to said providing, receiving, from said machine learning system, information regarding anti-noise to be generated by an acoustic transducer to compensate for noise under said at least one value, resulting from the operation of said magnetic resonance imaging system using a first set of said gradient coil pulse commands, experienced by the ears of the subject; Controlling said first magnetic resonance imaging system using said pulse sequence command and a first set of said gradient coil pulse commands to obtain said imaging k-space data; Operating said acoustic transducer to generate anti-noise using the information output by said trained machine learning system synchronized with said control of said magnetic field gradient coil system using the selected said gradient coil pulse command; A processor for controlling said magnetic resonance imaging system to perform the above; A magnetic resonance imaging system comprising the above.

10. The magnetic resonance imaging system according to claim 9, wherein at least one of said at least one value refers to a preparation scan command.

11. The magnetic resonance imaging system according to claim 9 or 10, wherein, by execution of said machine-executable instructions, said processor reconstructs magnetic resonance imaging data from said imaging k-space data.

12. The magnetic resonance imaging system according to any one of claims 9 to 10, wherein said trained machine learning system is implemented by a neural network that is part of said magnetic resonance imaging system.

13. The magnetic resonance imaging system according to any one of claims 9 to 10, wherein said magnetic resonance imaging system is coupled to a neural network that implements said trained machine learning system.

14. A memory storing machine-executable instructions; A medical system including a processor for controlling the medical system, By executing the machine-executable instructions, the processor is configured to respond to an input of a selected gradient coil pulse command for the magnetic resonance imaging system and generate anti-noise information to be generated by an acoustic transducer of the magnetic resonance imaging system, and at least one further parameter describing the imaging of the training subject having noise experienced by the training subject within the magnetic resonance imaging system. Control the medical system to provide a trained machine learning module that outputs at least one value related to the magnetic resonance imaging system, Providing the trained machine learning module includes Preparing a machine learning model, A training set of gradient coil pulse commands and a combination of the training values related to the parameters describing the imaging of the training subject having noise generated over time when operating a training magnetic resonance imaging system using the training set of gradient coil pulse commands experienced by the training subject. Providing training data including Training the machine learning model using the training set of gradient coil pulse commands, the training values as inputs, and the noise as outputs, wherein the training results in the trained machine learning system. Training, including A medical system.

15. A computer program including machine-executable instructions for controlling a magnetic resonance imaging system to acquire imaging k-space data from an imaging zone, wherein the magnetic resonance imaging system includes a magnetic field gradient coil system for generating a magnetic gradient field within the imaging zone, and a memory including pulse sequence commands for controlling the first magnetic resonance imaging system to acquire the imaging k-space data according to a magnetic resonance imaging protocol, and the memory further includes gradient coil pulse commands for controlling the magnetic field gradient coil system during the acquisition of the imaging k-space data. Upon execution of the machine-executable instructions, a processor receives a selection input for selecting a first set of gradient coil pulse commands from the gradient coil pulse commands, provides the selected gradient coil pulse commands and at least one value associated with each of at least one further parameter describing imaging of the subject to a machine learning system, receives, in response to the providing, information regarding anti-noise to be generated by an acoustic transducer to compensate for noise experienced in the subject's ear and resulting from the operation of the magnetic resonance imaging system using the first set of gradient coil pulse commands, under the at least one value, wherein the machine learning model is trained using a training set of gradient coil pulse commands, training values as inputs, and the noise as outputs, controls the first magnetic resonance imaging system using the pulse sequence commands for acquiring the imaging k-space data and the first set of gradient coil pulse commands, operates the acoustic transducer to generate anti-noise using the information output by the trained machine learning system synchronized with the control of the magnetic field gradient coil system using the selected gradient coil pulse commands A computer program for controlling the magnetic resonance imaging system to perform

Images