Determining 3D positioning data in an MRI system

Abstract

The present invention provides a means for determining 3D position data within an MRI system. The method for motion correction of MR data comprises the steps of generating (S10) a 3D model of a region of interest 24 of the object 23 by a computing unit 51, the region of interest 24 including at least one landmark 27 specific to the object 23, acquiring (S20) a 2D image of at least a part of the object 23 in an MRI system 22 by a first measuring device 20, 25, 52, the measuring device being disposed in a bore of the MRI system, determining (S30) by a computing unit 53 at least one landmark 27 in the 2D image, the at least one landmark 27 in the 2D image corresponding to at least one landmark 27 of the 3D model, determining (S40) by a computing unit 54 a 3D position of the region of interest 24 of the object 23 in the MRI system 22 based on the determined at least one landmark 27 in the 2D image, and providing (S50) by a computing unit 55 the 3D position of the region of interest 24 of the object 23 for motion correction of the MR data.

Description

[Technical field]

[0001] The present invention relates to a method for motion correction of MR data, an apparatus for motion correction of MR data, a system for medical imaging and a computer program element. [Background technology]

[0002] Magnetic Resonance Imaging (MRI) is known from state of the art. MRI is used to obtain medical images of anatomical structures, such as organs in the human body. MRI uses a strong magnetic field, a magnetic field gradient, and radio waves to obtain medical images. The quality of the medical image is important for comparison with reference images and for the analysis of the medical image (e.g. determining specific areas within an organ). The quality of the medical image depends in particular on the alignment of the desired object to be imaged with the imaging system. The alignment information may be inaccurate or may change over time, which may negatively affect the quality of the resulting medical image.

[0003] The paper by A. Kyme et al., “Marker-free optical stereo motion tracking for in-bore MRI and PET-MRI application,” Medical Physics, Vol. 47, No. 8 (June 1, 2020), discloses a method for prospective motion correction.

[0004] A paper by A. Kyme et al. entitled "Markerless motion tracking of awake animals in positron emission tomography" (IEEE transactions on medical imaging, Vol. 33, No. 11 (November 1, 2014)) discloses a method for motion compensation in positron emission tomography.

[0005] US Patent Application Publication No. 2018 / 0325415A1 discloses a method for measuring motion information from a human or animal subject during a medical imaging examination. Summary of the Invention

[0006] There is therefore a need for improved object localization in MRI systems.The object of the present invention is achieved by the subject matter of the independent claims, further embodiments being incorporated in the dependent claims.

[0007] According to a first aspect, a method for motion correction of MR data is provided. The method comprises generating, by a computing unit, a three-dimensional (3D) model of a region of interest of a subject, the 3D model including at least one landmark specific to the subject. The method further comprises acquiring, by a first measuring device, a two-dimensional (2D) image of at least a portion of the subject in an MRI system, the measuring device being at least partially disposed in a bore of the MRI system. The method further comprises determining, by the computing unit, at least one landmark in the 2D image, the at least one landmark in the 2D image corresponding to at least one landmark of the 3D model. The method further comprises determining, by the computing unit, a position of the region of interest of the subject in the MRI system based on the determined at least one landmark in the 2D image, and providing, by the computing unit, a 3D position of the region of interest of the subject for motion correction of the MR data.

[0008] As used herein, the term MR data should be understood broadly and relates to data associated with an MRI procedure. The MR data may relate to data used to control an MRI system during a preparatory phase of an MRI procedure and / or during an operational phase of an MRI procedure. The MR data may include data and / or information used to generate adaptive and / or control signals for magnetic resonance gradients, radio frequency pulses, and receiver frequencies in the scanner of an MRI system. Motion correction of MR data during the preparatory and / or operational phases of an MRI procedure is also known as prospective motion correction. Thus, the preparatory phase precedes the operational phase. The MR data may relate to MR images acquired (e.g., collected) during an MRI procedure. Motion correction of MR images during a post-processing phase of an MRI procedure is also known as retrospective motion correction.

[0009] As used herein, the term computing unit should be understood in a broad sense and refers in particular to a unit configured to process data in order to determine one or more landmarks in a 2D image by data processing. The computing unit may be a hardware unit (such as a controller, a workstation, a server), a software unit (such as a virtual machine running on a hardware unit), or a combination of hardware and software. The control unit may be in a single entity or distributed among several entities. The entities may be hardware and / or software units.

[0010] As used herein, the term 3D model should be understood broadly and relates to a model configured to describe a region of interest of a subject. The 3D model may be a point model, a line model, a surface model, or a volume model. For example, the 3D model may include facial features, bones, tissues, organs, and / or veins, and may also incorporate other body parts, organs, etc. The 3D model may be based on statistical information from a database, etc., taking into account, for example, age, sex, weight, etc. For example, the 3D model may be based on an anatomical atlas. The 3D model may also be based on past data of the subject to be imaged (such as previous images from a previous medical imaging examination). The 3D model may be described, for example, by a vector, which may include 3D information of at least one landmark, preferably several landmarks.

[0011] As used herein, the term subject means a human or animal. As used herein, an area of ​​interest relates to any part of a human or animal body (such as a bone, tissue, organ, or a combination thereof, although other body parts, organs, etc. may also be captured as information).

[0012] As used herein, the term landmark should be understood broadly and relates to at least one marker configured to be determined in an image including 2D images and 3D data. The landmark may be a structural (e.g., intrinsic) component of the region of interest (such as a bone, an eye, a nose, a hand, etc.), a color combination or color change (such as a mole on the left cheek), or a virtual marker adjacent to the region of interest (derived from two or more physical markers of the region of interest, such as two bones). A landmark is not a separate entity specifically attached to the object, but rather is intrinsic to the object. A landmark may be a part of the object or an adjacent region that is visible in the 2D image. The landmark may be determined by an image analysis algorithm, for example an edge detection algorithm. The landmark may be automatically determined in the 2D image by an image analysis algorithm. The 2D images are continuously acquired by a video stream from an in-bore camera. In particular, the at least one landmark in the 2D image may be acquired by searching for at least one landmark in the 3D model. As an example, the image analysis algorithm specifically scans the 2D image for specific features of at least one landmark in the 3D model. As another example, the image analysis algorithm can find landmarks in the 2D images and match them with at least one landmark in the 3D model. The landmarks are preferably predefined. If multiple landmarks (e.g., five) are predefined and only four landmarks are determined within the bore of the MRI system, the method can continue to run with the four landmarks.

[0013] As used herein, the term "first measuring device" should be understood broadly and relates to any measuring device configured to acquire a 2D image of a part of an object in the bore of an MRI system. The first measuring device may be a sensor unit. The measuring device may be an optical camera sensor, an infrared sensor, a laser interferometer, etc. The first measuring device may be a single entity or distributed in two or more entities. The entity is related to the sensor unit. The first measuring device may be, for example, an RGB sensor or a CCD sensor. The first measuring device may be an infrared camera, etc., streaming continuously from within the bore. The first measuring device may acquire the 2D image continuously or after a discrete time. The first measuring device may be arranged to be in direct visual contact with the object, in particular with a part of the object, more specifically with a region of interest of the object. The first measuring device may be arranged to be in indirect visual contact with the object, in particular with a part of the object, more specifically with a region of interest of the object by means of a mirror. The first measuring device may be in a wired connection (e.g. Ethernet) with a computing unit, data storage, server, workstation. The first measuring device may be in a wireless connection (e.g. WIFI) with the aforementioned entities.

[0014] As used herein, an MRI system refers to a modern MRI system configured to perform an MRI procedure. As used herein, an MRI system includes at least an MRI control configured to control the MRI system, a bore in which a movable support structure for supporting a subject is disposed, one or more MR source coils, and one or more MR detector coils. The MRI system may be advantageously augmented by a first measuring device.

[0015] As used herein, the term 3D position refers to at least three translational coordinates (such as x, y, z coordinates) of a single point of a region of interest of a subject. The 3D position of one or more single points may reveal the orientation of the region of interest of the subject. The 3D position may include three additional rotational coordinates.

[0016] The invention is based on the discovery that knowledge of the 3D position of the subject relative to the MRI system in an MRI procedure, and in particular knowledge of the position of the subject's region of interest (e.g. expressed in MRI system coordinates), is critical to the quality of the resulting MR images. To acquire MR images, the measurement data resulting from the MRI procedure must be further processed. The further processing to acquire MR images requires the 3D position of the subject's region of interest. If the used 3D position of the subject's region of interest in the further processing to acquire MR images is inaccurate, the resulting MR images will also be inaccurate. The 3D position of the subject's region of interest may change over time as the subject breathes in and out or simply moves. However, changes in the 3D position of the region of interest will affect the metadata of the MRI procedure and thus the corresponding acquired MR images if no motion correction is performed. The invention determines the 3D position of the subject's region of interest in the bore of the MRI system and uses the 3D position of the subject's region of interest to correct the MR images after imaging (i.e. retrospective motion correction). Furthermore, the determined 3D position of the region of interest of the subject can also be used to adapt the magnetic resonance gradients, radio frequency pulses and receiver frequency of the MRI system prior to the MRI procedure, in particular the MRI procedure (i.e. prospective motion correction). The determination of the 3D position of the region of interest of the subject is performed using 2D images acquired by modern measuring devices, in particular 2D camera sensors implemented in the bore of the MRI system. This can be advantageous since only one 2D camera is required, in particular no 3D depth camera is required. The 2D camera can advantageously be operated in the magnetic field in the bore of the MRI system compared to a 3D camera (such as a depth camera). The invention allows the use of simple but robust 2D cameras by mapping information from the acquired 2D images to a three-dimensional model, thereby determining the 3D position of the region of interest of the subject. The invention allows the correction of magnetic resonance motion artifacts using a single in-bore camera and an additional depth camera placed outside the bore (e.g. inside the scanner room).

[0017] In an embodiment, the method further comprises acquiring at least one modeling image of the object by a second measuring device placed outside the bore of the medical imaging system, the at least one modeling image being used to generate a 3D model of the region of interest of the object. As used herein, the term modeling image means that the modeling is simply used to generate a 3D model of the region of interest of the object in preparation for an MRI procedure. For example, the object lies on a support structure outside the bore of the MRI system, and the second measuring device is placed above the object (e.g. on the ceiling). The measuring device may take one or more modeling images of the region of interest from one or more viewpoints. The 3D model may include a default 3D model with default dimensions, these dimensions being adapted by the at least one modeling image of the region of interest of the object. The actual dimensions of the region of interest of the object can be obtained directly from the modeling image by analyzing the modeling image. If the modeling image only contains 2D data, this analysis may include utilizing neural networks or the like. If the modeling image contains 3D data, this analysis may simply include reading from the modeling image. The 3D model is, for example, newly generated during the implementation of the method described herein. That is, there is no default 3D model. In short, this may be advantageous since the accuracy of the 3D model and therefore the resulting quality of the MR data is increased. The modeling images may be continuously received from the second measuring device. The modeling images may be continuously analyzed to detect at least one landmark. The analysis may then include one or more mathematical sub-algorithms. The at least one landmark may be predefined for the region of interest (such as the cheekbone of the head). The mathematical algorithm may include an edge detection algorithm, a neural network trained for this purpose, or other suitable calculation methods. The mathematical algorithm may determine whether the subject is on the bed of the MRI system and / or whether the subject is outside or inside the bore of the MRI system. The mathematical algorithm may determine different parts of the subject (such as the head, legs, arms, etc.).The mathematical algorithm may use any data stream from the second measurement device, such as 2D information (RGB output from the depth camera), 3D information from the depth camera, or a combination thereof. The mathematical algorithm may first detect a predefined region of interest (such as the head) and then detect at least one predefined landmark (such as the cheekbone). If multiple modeling images are acquired, generating the 3D model of the region of interest of the subject may include calculating an average of the 3D models. If multiple modeling images are acquired, generating the 3D model of the region of interest of the subject may include fusing different 3D models of the region of interest of the subject.

[0018] In an embodiment, at least one modeling image may include 3D data of the region of interest of the subject. The 3D data may include translation coordinates (x, y, z directions) of each pixel in the modeling image, etc. The 3D data may be obtained from a depth camera, a laser interferometer scanner, and / or two or more 2D cameras (such as two RGB sensor cameras at different viewpoints that allow computer stereo vision). Computer stereo vision is the extraction of 3D information from a digital image. The 3D data may be extracted by examining the relative positions of objects (such as landmarks) in the digital image by comparing information from the region of interest from two viewpoints. The 3D data of the region of interest of the subject may advantageously increase the accuracy of the 3D model and therefore the resulting MR data. The 3D data from the modeling image may be used to obtain the corresponding 3D spatial positions in an appropriate coordinate system (such as an MRI coordinate system, a region of interest (such as a head) coordinate system, etc.). For example, a patient-centered coordinate system may be used. In this case, at least one landmark is the origin of the coordinate system. The coordinate system may include homogeneous coordinates to simplify calculations.

[0019] In an embodiment, the second measuring device can be a depth camera. The depth camera can advantageously provide highly accurate 3D information data of the object's region of interest. The depth camera can determine the 3D information from the 2D image using the time-of-flight principle. The depth camera can determine the 3D information from the 2D image using structured light. The depth camera can use coherent light and measure the phase shift of the reflected light relative to the source light (i.e., laser interferometry).

[0020] In an embodiment, the second measurement device may be at least one optical camera. The optical camera may be an RGB camera. The RGB camera provides only 2D images. Therefore, to obtain 3D data, either two images of two different viewpoints from an RGB camera are required (i.e., computer stereo vision, which preferably requires at least two optical cameras), or the 2D data from the 2D images acquired by the RGB camera must be aligned with the 3D model. The second option may be performed by a mathematical algorithm. The mathematical algorithm may be trained to process one or more inputs into one or more outputs by an internal processing chain, typically having a set of free parameters. The internal processing chain may be organized in interconnected layers that are traversed in succession in going from input to output. The mathematical algorithm may be trained by using a record of training data. The record of training data includes training input data and corresponding training output data. As used herein, the training input data may be a 2D image from a region of interest of the subject, and the training output data may be 3D data of the region of interest of the subject (e.g., measured with a 3D depth camera). The training input data and the training output data may also be simulated data to reduce the effort of providing training data. In short, this may be advantageous in terms of cost reduction and accuracy of the 3D model. The optical camera may be an infrared camera. The optical camera may use a charge-coupled device (CCD) sensor or an active pixel sensor (i.e., a complementary metal oxide semiconductor (CMOS) sensor).

[0021] In an embodiment, generating the 3D model may be based on a machine learning system representing a mathematical algorithm that processes at least one landmark of the region of interest of the subject. The machine learning system is trained to describe a relationship between geometric data of the region of interest of the subject and at least one landmark of the region of interest of the subject. The training data may be obtained from recordings of 2D images and corresponding 3D images. The 2D images and corresponding 3D images may be simulated to reduce the effort of providing the training data. The machine learning system may be implemented by a neural network, a machine learning algorithm, a convolutional neural network, or a generative adversarial network. Using a machine learning system may be advantageous in terms of efficiency and accuracy of the generation of the 3D model.

[0022] In an embodiment, the determination of the position of the region of interest of the subject in the MRI system may include determining one or more rotations and one or more translations of the 3D model relative to the position of the first measuring device in order to obtain a projection of the 3D model that fits the 2D image acquired by the first measuring device in the bore. In other words, the method checks which 3D positions of the region of interest of the subject can lead to the acquired projection (i.e., 2D image) of the region of interest of the subject. Thus, the determination may further take into account the position of the first measuring device in the bore. The determination may calculate, in reference, a viewpoint from the position of the first measuring device in the bore. The determination may use the physical equation of the intercept theorem. The determination may use a numerical approximation method to determine the position of the region of interest of the subject in the MRI system. The determination may further use a neural network, possibly trained for this purpose, to determine the position of the region of interest of the subject in the MRI system. In short, this may be advantageous to accurately determine the position of the subject in the MRI system.

[0023] In an embodiment, the first measuring device may be placed in the coil, or its housing, etc. Placing the first measuring device in the coil may be advantageous since nothing is hidden or the view between the measuring device and the region of interest of the subject is not blocked. The coil may be a head coil or another body coil. The coil may be a fixed coil in the MRI system. The first measuring device may also be attached to the ceiling of the bore. The first measuring device may also change its viewpoint depending on the subject and / or the region of interest (e.g., objects of different sizes require different viewpoints). The measuring device may also be placed in combination with one or more mirrors to image hidden areas (such as under the jaw). If mirrors are used, the position of the mirrors in addition to the position of the measuring device is used to determine the position of the subject in the MRI system.

[0024] In an embodiment, the modeling images can be used to obtain the corresponding 3D positions in an appropriate coordinate system. For example, a patient-centered coordinate system can be used. In this case, one of several predefined landmarks is used as the origin. Typically, multiple modeling images of the region of interest are available during the examination preparation, so the resulting 3D model can be averaged and stitched for accuracy. Using homogeneous coordinates, the resulting individual reference landmark positions can be expressed as vectors.

number

number

number

number

number

number

number

number

[0025] In an embodiment, the first measuring device may be an optical camera. The optical camera may be an RGB camera. The RGB camera is a measuring device that has a robust operating behavior in a magnetic environment, such as an MRT system. This may be advantageous in terms of robustness and reliability of the method. The optical camera may be an infrared camera. This may also be advantageous in terms of robustness and reliability of the method. It is mentioned that using a single optical camera as the first measuring device is sufficient to perform the method. However, it may also be advantageous to use many RGB cameras in the bore to obtain a better coverage of the region of interest of the subject in the MRI system.

[0026] In an embodiment, the region of interest of the subject may be the subject's head.

[0027] In an embodiment, the step of generating a 3D model of the object may be based on an anatomical model, preferably a morphable 3D model. This may be advantageous in terms of the accuracy of the 3D model. The morphable 3D model may be advantageous when the region of interest is a head or face. The morphable 3D model is a generative model, which may be established on a set of face or head examples in a registration procedure. The morphable 3D model may be a statistical model of the distribution of face or head examples.

[0028] In an embodiment, the position of the region of interest of the subject can be continuously determined and provided for motion correction of the MR data. The continuous execution of the method may require real-time capable hardware components for calculations and data exchange. The calculations may be performed by workstations, FPGAs, high performance computers, data centers. The data exchange between the first measuring device, the second measuring device, the control part of the MRI system and the calculation unit may be performed by third generation bus systems, Ethernet and fast Ethernet hubs.

[0029] According to a further aspect, an apparatus for motion correction of MR data is provided. The apparatus comprises a generating unit configured to generate a 3D model of a region of interest of a subject including at least one landmark, an acquiring unit configured to acquire a 2D image of at least a part of the subject in an MRI system, the acquiring unit being disposed in a bore of the MRI system, a first determining unit configured to determine at least one landmark in the 2D image, the at least one landmark in the 2D image corresponding to at least one landmark of the 3D model, a second determining unit configured to determine a position of the region of interest of the subject in the MRI system based on the determined at least one landmark in the 2D image, and a providing unit configured to provide a position of the region of interest of the subject for motion correction of the MR data. The generating unit, the first determining unit, the second determining unit, and the providing unit may be separate hardware units or separate software units executing on one or more hardware units, or a combination thereof. The hardware unit may be a controller, a computer, a server, a workstation. Data exchange between one or more hardware units may be via wires (Ethernet, Profinet, etc.) or wireless (WIFI, WLAN, etc.).

[0030] According to a further aspect, a system for medical imaging is provided. The system includes the above-mentioned apparatus, an MRI system, a first camera configured to acquire 2D images within the MRI system, and optionally a second camera configured to acquire images outside the MRI system. The second camera may be a depth camera. The first camera may be an RGB camera.

[0031] According to a last aspect, a computer program element is provided, which is configured to execute the steps of the above method when executed by a processor. The processor may be part of the medical imaging system or may be provided separately in another computing device. The computer program element may be stored in a computing unit which may be part of an embodiment. This computing unit may be configured to execute or induce the execution of the steps of the above method. Furthermore, it may be configured to operate the components of the above device. The computing unit may be configured to operate automatically and / or to execute the instructions of a user. The computer program may be loaded into a working memory of a data processor. The data processor may thus be ready to execute a method according to one of the above embodiments. This exemplary embodiment of the invention covers both computer programs using the invention from the beginning and computer programs that convert existing programs into programs using the invention by updates. Furthermore, the computer program element may provide all the steps necessary to carry out the procedures of the above exemplary embodiment of the method. According to a further exemplary embodiment of the invention, a computer readable medium such as a CD-ROM, a USB stick, etc. is presented. On the computer readable medium, a computer program element is stored. This computer program element has been described in the previous section. The computer program may be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid-state medium, supplied 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 communication systems. However, the computer program may also be presented over a network, such as the World Wide Web, and may be downloaded from such a network into the working memory of a data processor.According to a further exemplary embodiment of the invention, a medium is provided making available a computer program element for downloading, said computer program element being configured to perform a method according to one of the aforementioned embodiments of the invention.

[0032] It should be noted that the above embodiments may be combined with each other regardless of the aspect they relate to. Thus, the method may be combined with structural features of devices and / or systems of other aspects, and similarly, the devices and systems may be combined with each other's features and with the features described above with respect to the method.

[0033] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter.

[0034] Exemplary embodiments of the invention are illustrated in the following drawings. [Brief description of the drawings]

[0035] [Figure 1] 1 is a schematic diagram of a portion of a medical imaging system according to an embodiment of the present disclosure. [Diagram 2] FIG. 2 is a schematic diagram of an apparatus according to a further embodiment of the present disclosure. [Diagram 3] FIG. 11 is a flow chart diagram of a method for motion correction of MR data according to a further embodiment of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0036] FIG. 1 shows a schematic diagram of a portion of a medical imaging system according to a first embodiment of the present disclosure.

[0037] The medical imaging system 10 comprises an apparatus 11 (see FIG. 2) for motion correction of MR data. The system 10 further comprises an MRI system 22. The MRI system 22 comprises an MRI controller 12, a bore 13, three magnetic coils 14, 15, 16, a movable bed 17, 18, a first camera 20, 25 and a second camera 21. The second camera 21 is configured to acquire images outside the MRI system. The second camera is in this case mounted on the ceiling above the movable bed 18. The second camera 21 is a depth camera capable of providing images with 3D information data of a region of interest 24 of an object 23. Alternatively, instead of a depth camera, a laser interferometer scanner can also be used to acquire 3D information data of a region of interest 24 of an object 23. Optionally, two or more 2D cameras (such as RGB cameras) can also be used to acquire 3D data of a region of interest 24 of an object 23. The object 23 is a human being in this embodiment. Optionally, the subject 23 may be an animal (such as a sheepdog). The region of interest 24 is the head of the subject 23 in this example. Optionally, any other region may be the region of interest, such as the stomach, chest, arms, legs, etc. The region of interest is the area that is imaged in the MRI procedure. The subject 23 lies on a movable bed outside the MRI system 22. The region of interest 24 includes one landmark 27, in this case the cheekbone. Optionally, the region of interest may include multiple landmarks. Optionally, the landmark may also be a mole, an ear, an eye, a nose, etc. on the face. Optionally, the landmark may be a virtual marker adjacent to the region of interest, which virtual landmark is derived from two or more physical landmarks of the region of interest (such as two cheekbones). The second camera 21 is in a wired connection (such as Ethernet) with the device 11. Using the wired connection, in particular control signals and image data (i.e. 2D data and / or 3D data) are exchanged. In preparation for the MRI procedure, one or more modeling images of the subject 23 are acquired by the second camera 21. The device 12 can control or initiate the imaging process of the second camera 21. Once the preparation is completed, the movable bed 17 on which the subject 23 lies is moved into the bore 13 of the MRI system 22.Optionally, a head coil 19 is disposed on the subject's head 24. Inside the bore 13, three source and detector magnetic coils 14, 15, 16 are disposed on the inner surface of the bore 13. In this embodiment, inside the bore 13, a first camera 25 is disposed in the head coil 19. The first camera 25 is configured to acquire a 2D image of the region of interest 24, in particular the subject's head, inside the MRI system. The first camera 25 may be equipped with a mirror 26 disposed adjacent to the subject's head 24 for imaging hidden regions of the region of interest 24. Optionally, a first camera 20 may be disposed on the inner surface of the bore 13. Note that only one first camera is required. However, optionally, more than one first camera may be disposed in the bore to cover the entire region within the bore. The first cameras 20, 25 are 2D RGB cameras that generate 2D images. The first camera is in a wired or wireless connection, preferably a wired connection, with the device 11. The device 12 can control or initiate the imaging process of the first camera 20, 25. The device 12 can be in a wired connection with the MRI control unit 12 for exchanging data.

[0038] Fig. 2 shows a schematic diagram of an apparatus 50 according to a further embodiment of the present disclosure. The apparatus 50 is configured for motion correction of MR data of an MRI system. The apparatus 50 comprises a generation unit 51. The generation unit 51 is configured to generate a 3D model of a region of interest of a subject, comprising at least one landmark. The generation unit is in this example a software unit implemented in a hardware unit, which in this case is the CPU of a workstation. The generation unit 51 provides an anatomical model, in particular a morphable 3D model. The generation unit 51 receives from the second camera 21 one or more modeling images of a region of interest 24 of the subject 23, comprising the landmarks 27. The data exchange between the generation unit 51 and the second camera 21 is established by an Ethernet connection. The generation unit 51 processes the one or more modeling images. Here, processing means determining the positions of the landmarks and determining the size of the region of interest 24 of the subject 27 in order to generate a 3D model of the subject 27. The apparatus 50 further comprises an acquisition unit 52. The acquisition unit 52 is configured to acquire a 2D image of at least a part of the object in the MRI system 22, the acquisition unit 52 being arranged in the bore of the MRI system. The acquisition unit 52 is in this embodiment the first camera 20 or 25 (see description of FIG. 1). The device further comprises a first determination unit 53. The first determination unit 53 is configured to determine at least one landmark in the 2D image. The at least one landmark in the 2D image corresponds to at least one landmark of the 3D model. The first determination unit 53 is in this embodiment a software unit implemented in the same hardware unit as the generation unit 51. The software unit may include an image processing software module to analyze the 2D image. The first determination unit 53 transmits information of the determined at least one landmark to a second determination unit 54. The second determination unit 53 is part of the device 50. The second determination unit 54 is in this case a software unit implemented in the same hardware unit as the generation unit 51 and the first determination unit 53.The second determination unit 54 is configured to determine a position of the object in the MRI system based on the determined at least one landmark in the 2D image. The second determination unit 54 may include a mathematical algorithm trained to derive a 3D position of the object's region of interest from the 2D image. The mathematical algorithm is explained in more detail in FIG. 3. The apparatus 50 further comprises a providing unit 55, which in this embodiment is a software unit implemented in the same hardware unit as the generating unit 51. The providing unit 55 is configured to provide the position of the region of interest 24 of the object 27 for motion correction of the MR data. The providing unit 53 provides the position of the region of interest 24 of the object 27 to the MRI controller 12 for adjusting the settings of the MRI procedure or to a server (not shown) for correcting already acquired MRI images.

[0039] FIG. 3 shows a flow chart diagram of a method for motion correction of MR data according to a further embodiment. The method is composed of five steps. In a first step S10, a 3D model of a region of interest of a subject body is generated, comprising at least one landmark. The 3D model is generated by the aforementioned generating unit 51. Step S10 further comprises acquiring at least one modeling image of the subject. The at least one modeling image is acquired by a second camera 21, for example a depth camera. The at least one modeling image shows a region of interest of the subject, the modeling image being acquired outside the bore of the MRI system. The at least one modeling image comprises 3D data. The 3D data comprises translation coordinates (x, y, z directions) of each pixel in the modeling image, among others. The 3D model may be based on a machine learning system representing a mathematical algorithm that processes at least one landmark of the region of interest of the subject. The machine learning system is trained to describe a relationship between geometric data of the region of interest of the subject and the at least one landmark of the region of interest of the subject. The 3D model may further be an anatomical model, preferably a morphable 3D model. In step S20, a 2D image of at least a part of the object in the MRI system is acquired by a first measuring device 20, 25. The first measuring device is arranged in the bore of the MRI system. The first measuring device may be an RGB camera. The first measuring device may be arranged on the inner surface of the bore or in a coil, in particular a head coil. The first measuring device may be equipped with a mirror. The 2D images may be acquired continuously, for example every second, in order to detect movements of the region of interest of the object. In step S30, at least one landmark 27 in the 2D image is determined. The at least one landmark in the 2D image corresponds to at least one landmark of the 3D model. The determination of the at least one landmark is performed by an image analysis algorithm (such as edge detection). The image analysis algorithm may be trained using records of previous examinations. The records may contain multiple landmarks (special bones, nose, eyes, etc.).The image analysis algorithm receives from the generating unit 51 information of at least one landmark that has to be located. In step S40, a position of the region of interest of the object in the MRI system is determined based on the determined at least one landmark in the 2D image. This determination is performed by a second determining unit 54. The determination of the position of the object in the MRI system includes determining one or more rotations and one or more translations of the 3D model relative to the position of the first measuring device in order to obtain a projection of the 3D model that fits the 2D image acquired by the first measuring device in the bore. This determination may further take into account the position of the first measuring device in the bore. This determination may calculate a viewpoint from the position of the first measuring device in the bore. The determination may use the physics equation of the intercept theorem. The determination may use a numerical approximation method to determine the position of the region of interest of the object in the MRI system. The determination may further use a neural network to determine the position of the region of interest of the object in the MRI system. In step S50, a 3D position of the region of interest of the object for motion correction of the MR data is provided. The 3D positions may be transmitted to an MRI control of the MRI system for correcting magnetic resonance gradients, radio frequency pulses, receiver frequencies for future acquired MRI images, or to a server or the like for modifying already acquired MRI images. The positions of the subject's regions of interest may be continually determined and provided for motion correction of the MR data, which may include prospective and / or retrospective motion correction, as described above.

[0040] In another exemplary embodiment, a computer program or a computer program element is provided that is configured to carry out, on a suitable device or system, the steps of the method according to one of the above embodiments.

[0041] Thus, a computer program element may be stored in a data processing unit which may be part of the embodiments. This data processing unit may be configured to execute or direct the execution of the steps of the above-mentioned methods. Furthermore, the data processing unit may be configured to operate the components of the above-mentioned devices and / or systems. The computing unit may be configured to operate automatically and / or to execute the instructions of a user. The computer program may be loaded into a working memory of a data processor. The data processor may thus be equipped to execute a method according to one of the above-mentioned embodiments.

[0042] Moreover, the computer program element may provide all the steps required to carry out the procedures of the exemplary embodiments of the methods described above.

[0043] According to a further exemplary embodiment of the present invention, a computer readable medium such as a CD-ROM, a USB stick or the like is presented, on which computer program elements are stored, the computer program elements being as described in the previous section.

[0044] The computer program may be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid-state medium, supplied 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 communication systems.

[0045] However, the computer program may also be presented via a network such as the World Wide Web and downloaded from such a network into the working memory of a data processor. According to a further exemplary embodiment of the invention, a medium is provided making available a computer program element for downloading, the computer program element being configured to perform a method according to one of the aforementioned embodiments of the invention.

[0046] It should be noted that the embodiments of the present disclosure are described with reference to different subject matters. In particular, some embodiments are described with reference to method type claims, while other embodiments are described with reference to device type claims. However, a person skilled in the art may infer from the above and following descriptions that, unless otherwise specified, any combination of features belonging to one type of subject matter, as well as any combination of features related to different subject matters, are considered to be disclosed in the present application. However, all features may be combined if they provide a synergistic effect that is more than a mere collection of features.

[0047] While the invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the dependent claims.

[0048] In the claims, the word "comprising" does not exclude other elements or steps, and the singular elements do not exclude a plurality. A single processor or other unit may fulfill the functions of several items recited in the claims. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be interpreted as limiting the scope. Explanation of symbols

[0049] 10. System 11 Equipment 12 MRI Controller 13 Boa 14, 15, 16 coils 17, 18 Adjustable bed 19 Head coil 20, 25 First Camera 21 Second Camera 22 MRI System 23 Target 24 Areas of Interest 26. Mirror 27 Landmark 50 equipment 51 Generating Units 52 Acquired Units 53 First Decision Unit 54 Second Decision Unit 55 units offered Generate S10 3D model S20 Capture 2D Images S30 Determine at least one landmark Determine the 3D location of the region of interest within the S40 MRI system Provides 3D location of region of interest for motion correction of S50 MR data

Claims

1. Generating, by a computing unit, a 3D model of a target region of interest, the 3D model including at least one landmark unique to the target; Acquiring, by a first measurement device, at least one 2D image of at least a part of the target within an MRI system, the measurement device being at least partially disposed within a bore of the MRI system; Determining, by the computing unit, at least one landmark within the 2D image, the at least one landmark within the 2D image corresponding to the at least one landmark of the 3D model; Determining, by the computing unit, a 3D position of the target region of interest within the MRI system based on the at least one landmark determined within the 2D image; Providing, by the computing unit, the 3D position of the target region of interest for motion correction of MR data; In a method for motion correction of MR data, the method comprising: Further comprising acquiring at least one modeling image of the target, the at least one modeling image being at least one image acquired by a second measurement device disposed outside a bore of the MRI system, the at least one modeling image being used to generate the 3D model of the target region of interest.

2. The method according to claim 1, wherein the at least one modeling image includes 3D data of the target region of interest.

3. The method according to claim 1 or 2, wherein the second measurement device is a depth camera.

4. The method according to claim 1, wherein the second measurement device is at least two optical cameras.

5. The second measurement device is one optical camera, and the step of generating the 3D model of the target region of interest includes adjusting the modeling image of the target with a default 3D model. The method according to claim 1.

6. The step of generating the 3D model is based on a machine learning system representing a mathematical algorithm for processing at least one landmark of the region of interest of the subject, and the machine learning system is trained to describe the relationship between the geometric data of the region of interest of the subject and at least one landmark of the region of interest of the subject. The method according to any one of claims 1 to 5.

7. The step of determining the 3D position of the region of interest of the subject within the MRI system includes determining one or more rotations and one or more translations of the 3D model related to the position of the first measurement device in order to obtain a projection of the 3D model that fits the 2D image acquired by the first measurement device within the bore. The method according to any one of claims 1 to 6.

8. The first measurement device is arranged within a coil. The method according to any one of claims 1 to 7.

9. The first measurement device is an optical camera. The method according to any one of claims 1 to 8.

10. The region of interest of the subject is the head of the subject. The method according to any one of claims 1 to 9.

11. The step of generating the 3D model of the subject is based on an anatomical model, preferably a morphable 3D model. The method according to claim 1.

12. The 3D position of the region of interest of the subject is continuously determined and provided for the motion correction of the MR data. The method according to any one of claims 1 to 11.

13. A generation unit for generating a 3D model of a subject, wherein the 3D model includes at least one landmark specific to the subject, a generation unit, An acquisition unit for acquiring at least a part of 2D images of the subject within an MRI system, wherein the acquisition unit is arranged within the bore of the MRI system, an acquisition unit, A first determination unit for determining at least one landmark within the 2D image, wherein the at least one landmark within the 2D image corresponds to the at least one landmark of the 3D model, a first determination unit, A second determination unit for determining the position of the subject within the MRI system based on the determined at least one landmark within the 2D image. For motion correction of MR data, a providing unit that provides the position of the object and In an apparatus for motion correction of MR data, comprising: The apparatus further comprises a second measurement device disposed outside the bore of the MRI system that acquires at least one modeling image of the object. The generating unit generates the 3D model of the object using the at least one modeling image. An apparatus characterized by this.

14. The apparatus according to claim 13, and An MRI system, and A first camera that acquires a 2D image within the MRI system A system for medical imaging, comprising:

15. A computer program that, when executed by a processor, performs the steps of the method according to any one of claims 1 to 12.

Images