Model-based restoration of three-dimensional tomographic medical imaging data.

Abstract

Disclosed herein is a medical system 100, 300, 400 comprising a memory 110 storing machine-executable instructions 120 and a computing system 104. Execution of the machine-executable instructions causes the computing system to receive (200) corrected three-dimensional tomographic medical image data 122 in which facial features have been erased within a facial zone 128; fit (202) a model-based segmentation 126 to the corrected three-dimensional tomographic medical image data, the model-based segmentation comprising a plurality of surface meshes 602, 604, 606, 608, 700 defining an anatomical region 800; assign (204) an anatomical label to each of the voxels of a reconstructed region 130 of the corrected three-dimensional tomographic medical image data according to the anatomical region defined by the plurality of surface meshes; and assign (206) contrast values ​​to the voxels of the reconstructed region using the anatomical labels to provide a restored three-dimensional tomographic medical image data 132.

Description

[Technical field]

[0001] The present invention relates to tomographic medical imaging data, and in particular to the restoration of three-dimensional tomographic medical image data. [Background technology]

[0002] Various tomographic medical imaging techniques, such as magnetic resonance imaging (MRI) and computed tomography, allow detailed visualization of a subject's anatomy. It is a current trend to modify face regions in tomographic medical imaging data to make them unidentifiable before storing them in medical databases, or even before storing them in sets of data for training and testing purposes. The drawback is that data modified in this way is unlikely to work with many of the libraries of numerical algorithms for automatically processing and analyzing this data.

[0003] International Patent Application Publication WO2020 / 198560A1 discloses that facial information in medical images is de-identified using an approach that reduces the undesirable effects of de-identification on image statistics while protecting the privacy of participants. Rather than removing or blurring facial voxels, the facial voxels are replaced with image data of a template face, potentially an average face in a population. The participant's facial features are completely removed, but the output image is generated to resemble a complete craniofacial image with similar statistical image texture characteristics as the original, thereby minimizing the impact on downstream biomarker measurements. Summary of the Invention

[0004] The invention provides a medical system, a computer program product and a method in the independent claims. Embodiments are set out in the dependent claims.

[0005] WO2020 / 198650 uses templated image data to replace portions of deleted or obscured data in facial zones. This document only discloses or shows the inpainting of two-dimensional slices of data. The templates disclosed in this document are not suitable for inpainting three-dimensional data sets. The templates may not fit the three-dimensional data set properly.

[0006] The embodiments provide a means for improving the quality of restored 3D tomographic medical image data by adapting a model-based segmentation to the corrected 3D tomographic medical image data. The model-based segmentation comprises a number of surface meshes that three-dimensionally define volumes corresponding to anatomical regions. These models can be adapted to 3D datasets of any shape. Furthermore, the model-based segmentation can flexibly adapt to variations and inconsistencies in size and / or shape of facial zones.

[0007] The model is then used to provide anatomical labels to the voxels of the modified 3D tomographic medical image data that are within the reconstructed region. The reconstructed region includes at least a facial zone. The anatomical labels are then used to reconstruct facial structures within the facial zone, resulting in a repair of the 3D tomographic medical image data. The facial features within the repaired facial zone are defined by a surface mesh of the model-based segmentation, and therefore no longer resemble the subject from whom the tomographic medical image data was obtained. However, the repaired tomographic medical image data can be used in numerical algorithms and calculations that depend on the facial structures present.

[0008] In one aspect, the present invention provides a medical system comprising a memory storing machine executable instructions and a computing system. This medical system embodiment may be incorporated into various types of systems and used in various contexts. In one example, the medical system is a cloud-based or remote-based server used to perform image reconstruction and image processing tasks. In another embodiment, the medical system is a computer system or workstation accessible to a radiologist or other medical professional. As yet another example, the medical system is incorporated into or comprises a medical imaging system, such as a CT or computed tomography system, or a magnetic resonance imaging or MRI system.

[0009] Execution of the machine executable instructions causes the computing system to receive modified 3D tomographic medical image data describing a subject's head. The modified 3D tomographic medical image data comprises voxels. The 3D tomographic medical image data has been facially erased within the facial zone. The term "modified 3D tomographic medical image data" is a label used to indicate or identify a particular 3D tomographic medical image data set or data. As used herein, facially erased encompasses either facial features being removed or blurred. The effect of this is that it becomes impossible or more difficult to identify the subject from the modified 3D tomographic medical image data. As used herein, the modified 3D tomographic medical image data describes the head but incorporates at least a field of view that includes the brain or the subject.

[0010] Execution of the machine executable instructions further causes the computing system to fit the model-based segmentation to the corrected three-dimensional tomographic medical image data. The model-based segmentation model is the entire head. The model-based segmentation comprises a number of surface meshes that define anatomical regions. As used herein, the entire head encompasses the region of the subject's head including at least the entire skull and surrounding soft tissue. The model-based segmentation defines these surface meshes that are used to divide the region of the head into respective anatomical regions. For example, various bone and tissue structures are identified by the surface meshes. In some embodiments, the model-based segmentation uses a balance between internal and external energies to bend or modify the surface mesh. If the face is erased from the facial zone, either by deleting or blurring voxels, this part of the face is excluded from the external constraints. For example, a model is fitted to the region of the subject's head outside the facial zone, and then the model within the facial zone can be easily fitted and fitted to match the rest of the model.

[0011] Execution of the machine executable instructions further causes the computing system to assign, to each of the voxels of the reconstructed region of the corrected 3D tomographic medical image data, an anatomical label according to the anatomical region defined by the plurality of surface meshes. The reconstructed region includes at least the facial zone. There are several different variants. In one example, the reconstructed region may be identical to the facial zone. In this case, only the erased or blurred regions are subsequently labeled. In another case, the reconstructed region is larger than the facial zone but smaller than the subject's entire head or field of view. This means, for example, that the area to be labeled extends beyond the facial zone. In another example, the reconstructed region is identical to the subject's entire head or field of view of the corrected 3D tomographic medical image data.

[0012] Execution of the machine executable instructions further causes the computing system to assign contrast values ​​to voxels in the reconstructed regions using the anatomical labels to provide restored 3D tomographic medical image data. As described below, there are several ways to assign contrast values ​​or values ​​within a voxel. This embodiment is beneficial because it allows a way to restore the corrected 3D tomographic medical image data. Currently, tomographic medical image data tends to have facial areas erased or blurred in the facial zone. This causes various technical problems. Well-known algorithms for segmenting the subject's head or especially the brain rely on anatomical markers or structures in the facial zone to function properly. Another technical difficulty is that the facially erased tomographic medical image data is not useful for training machine learning algorithms such as neural networks. By constructing restored 3D tomographic medical image data, this facially erased tomographic medical image data can be used in other algorithms or even for training machine learning modules.

[0013] One of the potential advantages of using model-based segmentation to fit the corrected 3D tomographic medical image data is that it is useful in a variety of situations and provides a better replacement of voxels of the facial zones and / or reconstruction regions. For example, the corrected 3D tomographic medical image data may be a 3D dataset or even a stack of 2D datasets used to form the 3D dataset. Model-based segmentation can fit into these different layers or granularities in the 3D data. This provides better results than using techniques such as, for example, using templates to replace parts of the facial zones.

[0014] Assigning contrast values ​​includes assigning intensities to voxels in the reconstructed regions using the anatomical labels to provide restored 3D tomographic medical image data, which can be accomplished, for example, by having a look-up table that obtains an intensity value depending on which anatomical region a voxel is in.

[0015] This can also be extended by normalizing the look-up table using regions of the rectified 3D tomographic medical image outside the facial zone before assigning contrast values ​​to voxels in the reconstructed region.

[0016] For example, model-based segmentation is adapted to the corrected 3D tomographic medical image data. The average values ​​of voxels within the anatomical region and outside the reconstructed region or facial zone are compared to the corresponding anatomical region in the lookup table. The average lookup table is then scaled or normalized to adjust it so that the average lookup table returns the average intensity of voxels within the anatomical region and outside the reconstructed region or facial zone.

[0017] In another embodiment, execution of the machine executable instructions further causes the computing system to use the restored three-dimensional tomographic medical image data in any one of a segmentation algorithm, an electroencephalography algorithm, a magnetoencephalography algorithm, a transcranial magnetic stimulation planning algorithm, a transcranial direct current stimulation planning algorithm, and a transcranial alternating current stimulation planning algorithm. The replacement of voxels in the face zone is particularly beneficial when using or planning algorithms that rely on electromagnetic computation.

[0018] In another embodiment, execution of the machine executable instructions further causes the computing system to use the model-based segmentation as input to an electroencephalogram algorithm, a magnetoencephalogram algorithm, a transcranial magnetic stimulation planning algorithm, a transcranial direct current stimulation planning algorithm, and a transcranial alternating current stimulation planning algorithm for electromagnetic calculations within the subject's head. This embodiment is particularly useful because the segmentation defines regions of different tissue types within the subject's head. Typically, finite element methods are used to perform the electromagnetic calculations. This allows, with knowledge of the segmentation, to perform the electromagnetic calculations using forward calculations that do not use finite elements. This significantly reduces the computation time required to apply any one of these algorithms.

[0019] In another embodiment, the execution of the machine executable instructions causes the computing system to assign contrast values ​​to voxels in the reconstructed region using a look-up table with an entry for each anatomical label. This embodiment is beneficial because it provides a very simple, effective and fast means of filling voxels in the reconstructed region. This is particularly useful when the reconstructed region is the same as or only slightly larger than a facial zone and does not encompass the subject's brain region. When using a look-up table, the resulting restored tomographic medical image may not look equivalent to a tomographic medical image that has not had facial features removed and restored, but it is very effective in various types of numerical algorithms. In particular, it works very well in segmentation algorithms, and also in electromagnetic calculations that solve equations using electromagnetic fields or finite element methods.

[0020] In another embodiment, the execution of the machine executable instructions further causes the computing system to normalize the lookup table with a region of the modified 3D tomographic medical image other than the face zone before assigning contrast values ​​to the voxels in the reconstructed region. Different CT, MRI or other tomographic imaging techniques have different contrasts from image to image. Therefore, normalizing the lookup table or adjusting the values ​​of the lookup table to match the modified 3D tomographic medical image data provides a more natural and effective result.

[0021] In another embodiment, execution of the machine executable instructions further causes the computing system to perform smoothing and / or contrast matching at the boundaries of the reconstructed region. This is beneficial because there is a slight discontinuity at the boundary between the reconstructed region and a portion of the corrected 3D tomographic medical image data that is outside the reconstructed region. Adding smoothing or contrast matching at the boundary provides the restored 3D tomographic medical image data that is more effective in numerical algorithms.

[0022] In another embodiment, the memory further comprises a reconstruction neural network configured to assign contrast values ​​to voxels of the reconstructed region in response to receiving the corrected three-dimensional tomographic medical image data and an anatomical label for each voxel of the reconstructed region. As used herein, the term "reconstruction neural network" is used to identify a particular neural network. The reconstruction neural network is a neural network. The reconstruction neural network can be formed, for example, from a pix2pix neural network or a CycleGAN neural network.

[0023] The reconstruction neural network is configured to take two things as inputs. One is the label for the anatomical value of the voxels in the reconstructed region, and the other is to take the corrected 3D tomographic medical image data as input. It is relatively easy to do this and build a neural network to receive both sets of data. It is fairly common to extend various neural network architectures to receive color data, for example data formatted in RGB format consisting of three numeric values. In this case, the neural network can be extended in the same manner, except that it is deployed to have two inputs. One input is the contrast or value of the voxels in the corrected 3D tomographic medical image data, and the other input is to receive a value that encodes the anatomical region.

[0024] For example, it is possible to use specific values ​​or ranges of values ​​to code different types of tissues or anatomical regions of the subject. In the reconstructed region, the arrangement of voxel values ​​can be done in different ways. In one example, the reconstruction neural network outputs the entire image, with the values ​​of the voxels in the reconstructed region replaced. This is similar to the idea of ​​an autoencoder, except that the neural network can replace all voxels with additional data from the anatomical data. In other cases, the output of the neural network is received, and then an additional algorithm obtains the values ​​of the voxels in the reconstructed region, which are then pasted into the corrected three-dimensional tomographic medical image data. In this way, voxels outside the reconstructed region are not replaced by the reconstruction neural network.

[0025] The reconstruction neural network can be trained in a variety of different ways. In one example, a complete three-dimensional dataset of tomographic medical image data can be obtained, from which training data can be prepared completely. In one example, the original three-dimensional dataset is used as ground truth data, and the input for supervised learning or deep learning is obtained by taking the same dataset and performing blurring or removal of those face zones. The original unmodified image can also be segmented with a model-based segmentation. A dataset containing the blurred or removed face regions together with the segmentation is then obtained, which can be input to the neural network and compared with the original dataset. Such datasets can be prepared in large quantities and used for supervised learning or deep learning.

[0026] There are also a variety of classic GAN approaches that could work. For example, pix-to-pix scanning, which converts a label mask into a photorealistic magnetic resonance image, works well. Another example is cycle GAN, which can be used, which uses unpaired images. There are literally a variety of ways that GAN neural networks can be used.

[0027] In another embodiment, the model-based segmentation comprises a defined brain volume. For example, the defined brain volume is a volume that outlines the border of the subject's brain. Execution of the machine-executable instructions causes the computing system to ensure that the brain volume of the restored 3D tomographic medical image data is identical to the brain volume of the corrected 3D tomographic medical image data. If the reconstructed region comprises all or part of the brain volume, the computing system obtains values ​​of original voxels in the brain volume from the corrected 3D tomographic medical image data. If the reconstructed region does not extend into the brain volume, no such processing is required.

[0028] In another embodiment, the corrected three-dimensional tomographic medical image data is a three-dimensional data set or a stack of two-dimensional slices.

[0029] In another embodiment, the modified three-dimensional tomographic medical image data is magnetic resonance image data or is computed tomography imaging data.

[0030] In another embodiment, the model-based segmentation is a deformable shape model segmentation.

[0031] In another embodiment, the model-based segmentation is locally deformable statistical shape model segmentation.

[0032] In another embodiment, the model-based segmentation is active contour model segmentation.

[0033] In another embodiment, execution of the machine executable instructions further causes the computing system to receive the acquired three-dimensional tomographic medical image data. Execution of the machine executable instructions further causes the computing system to receive modified three-dimensional tomographic medical image data in response to inputting the acquired three-dimensional tomographic medical image data into a face removal algorithm. As used herein, a face removal algorithm encompasses an algorithm that uses segmentation or registration techniques to determine a facial zone. The face removal algorithm takes voxels within this facial region and sets them to a predetermined value, such as zero, or blurs them by their neighbors to obscure facial features.

[0034] In another embodiment, the medical system further comprises a tomographic medical imaging system. Execution of the machine-executable instructions further causes the computing system to control the tomographic medical imaging system to obtain the acquired three-dimensional tomographic medical imaging data.

[0035] In another embodiment, voxels within the face zone are removed or blurred to eliminate the face.

[0036] In another embodiment, removing a voxel from the face zone comprises setting the voxel to a predetermined value or removing the value assignment.

[0037] In another embodiment, the facial zone comprises at least a portion of the facial bone structure.

[0038] In another embodiment, the reconstructed region is within an anatomical region. In another embodiment, each voxel of the reconstructed region of the corrected three-dimensional tomographic medical image data is assigned a corresponding anatomical label according to the anatomical region defined by the multiple surface meshes. In other words, the anatomical label for a particular voxel is determined by its anatomical region defined by the multiple surface meshes.

[0039] In another aspect, the present invention provides a method of medical imaging, comprising receiving modified three-dimensional tomographic medical imaging data describing a subject's head and comprising voxels, the three-dimensional tomographic medical imaging data being facially erased within a facial zone, the method further comprising fitting a model-based segmentation to the modified three-dimensional tomographic medical image data, the model-based segmentation model being the entire head, the model-based segmentation comprising a plurality of surface meshes defining anatomical regions.

[0040] The method further comprises the step of labelling each voxel of a reconstructed region of the corrected three-dimensional tomographic medical image data with an anatomical label according to an anatomical region defined by the plurality of surface meshes, the reconstructed region including at least a facial zone, and the method further comprises the step of assigning contrast values ​​to the voxels in the reconstructed region with the anatomical label to provide restored three-dimensional tomographic medical imaging data.

[0041] In another aspect, the present invention provides a computer program product including machine executable instructions for execution by a computing system, e.g., they are machine executable instructions on a non-transitory storage medium. Execution of the machine executable instructions causes the computing system to receive modified three-dimensional tomographic medical image data describing a subject's head and including voxels. The three-dimensional tomographic medical imaging data is facially erased within a facial zone. Execution of the machine executable instructions further causes the computing system to apply a model-based segmentation to the modified three-dimensional tomographic medical image data. The model-based segmentation model is an entire head. The model-based segmentation comprises a plurality of surface meshes that define anatomical regions.

[0042] Execution of the machine-executable instructions further causes the computing system to label each of the voxels of the reconstructed region of the corrected three-dimensional tomographic medical image data with an anatomical label according to an anatomical region defined by a plurality of surface meshes, the regions being regions defined by being between certain of the plurality of surface meshes. The reconstructed region includes at least a facial zone. Execution of the machine-executable instructions further causes the computing system to assign contrast values ​​to the voxels in the reconstructed region with the anatomical label to provide the restored three-dimensional tomographic medical image data.

[0043] It will 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.

[0044] As will be appreciated by those skilled in the art, aspects of the present invention may be embodied as an apparatus, method, or computer program product. Accordingly, aspects of the present 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 of which are generally referred to herein as "circuits," "modules," or "systems." Additionally, aspects of the present invention may take the form of a computer program product embodied in one or more computer-readable medium(s) having computer-executable code embodied thereon.

[0045] Any combination of one or more computer readable media may be utilized. The computer readable medium may be a computer readable signal medium or a computer readable storage medium. A "computer readable storage medium" as used herein encompasses any tangible storage medium that stores instructions executable by a processor or a computing system of a computing device. The computer readable storage medium is referred to as a non-transitory computer readable storage medium. The computer readable storage medium is also referred to as a tangible computer readable medium. In some embodiments, the computer readable storage medium may also store data accessible by the computing system 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 register files of a computing system. 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 accessible by a computer device over a network or communication link. For example, data may be obtained over a modem, over the Internet, or over a local area network. Computer executable code embodied in the computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, fiber optic cable, RF, etc., or any suitable combination of the foregoing.

[0046] A computer-readable signal medium includes a propagated data signal having computer-executable code embodied therein, 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, electromagnetic, 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 can communicate, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device.

[0047] "Computer memory" or "memory" is an example of a computer-readable storage medium. Computer memory is memory directly accessible to a computing system. "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.

[0048] A "computing system," as used herein, encompasses electronic components capable of executing programs or machine-executable instructions or computer-executable code. References to a computing system, including examples of a "computing system," should be interpreted as potentially including multiple computing systems or processing cores. A computing system is, for example, a multi-core processor. A computing system also refers to a collection of computing systems within a single computer system or a collection of computing systems distributed among multiple computer systems. The term computing system should also be interpreted as potentially referring to a collection or network of computing devices, each of which comprises a processor or computing system. Machine-executable code or instructions are executed by multiple computing systems or processors, which may be present within the same computing device or even distributed across multiple computing devices.

[0049] Machine-executable instructions or computer-executable code include instructions or programs that cause a processor or other computing system to perform aspects of the present invention. Computer-executable code that performs operations for 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 traditional 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 used with an interpreter that generates the machine-executable instructions on the fly. In other examples, the machine-executable instructions or computer-executable code is in the form of programming for a programmable logic gate array.

[0050] The computer executable code may run entirely on the user's computer as a stand-alone software package, partially on the user's computer, 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 via any type of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet Service Provider).

[0051] Aspects of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the present invention. It is understood that each block or part of blocks of the flowcharts, diagrams, and / or block diagrams, where applicable, can be implemented by computer program instructions in the form of computer executable code. It is further understood that combinations of blocks from different flowcharts, diagrams, and / or block diagrams can be obtained when not mutually exclusive. These computer program instructions are provided to a computing system of a general purpose computer, a special purpose computer, or another programmable data processing device to manufacture a machine, whereby the instructions, executed via the computing system of the computer or other programmable data processing device, create means for performing the functions / acts specified in the blocks of the flowcharts and / or block diagrams.

[0052] These machine-executable instructions or computer program instructions may also be stored on a computer-readable medium that can direct a computer, another programmable data processing apparatus, or another device to function in a particular manner, such that the instructions stored on the computer-readable medium produce an article of manufacture that includes instructions that perform the functions / acts specified in the flowchart and / or block diagram blocks.

[0053] The machine-executable instructions or computer program instructions may also be loaded into a computer, another programmable data processing apparatus, or another device to cause a sequence of operational steps to be executed on the computer, another programmable apparatus, or another device to produce a computer-implemented process, such that the instructions executing on the computer or another programmable apparatus provide a process for performing the functions / acts specified in the block blocks of the flowcharts and / or block diagrams, etc.

[0054] 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 allows a computer to receive input from an operator and provides output from the computer to a user. In other words, a user interface allows an operator to control or manipulate a computer, and the interface allows the computer to show the effects 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 for the reception of information or data from an operator.

[0055] "Hardware interface," as used herein, encompasses an interface that allows a computing system of a computer system to interact with and / or control external computing devices and / or devices. A hardware interface allows a computing system to send control signals or instructions to external computing devices and / or devices. A hardware interface also allows a computing system 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.

[0056] A "display" or "display device," as used herein, encompasses an output device or user interface suitable for displaying images or data. A display outputs visual, audio, and / or tactile data. Examples of displays include, but are not limited to, computer monitors, television screens, touch screens, tactile electronic displays, Braille screens, cathode ray tubes (CRTs), storage tubes, bi-stable displays, ePaper, vector displays, flat panel displays, vacuum fluorescent displays (VFs), light emitting diode (LED) displays, electroluminescent displays (ELDs), plasma display panels (PDPs), liquid crystal displays (LCDs), organic light emitting diode displays (OLEDs), projectors, and head mounted displays.

[0057] Three-dimensional tomographic medical image data is defined herein as a reconstructed three-dimensional visualization of anatomical data imaged by a tomographic medical imaging system, such as a computed tomography system or a magnetic resonance imaging system.

[0058] K-space data is defined herein as the recorded measurements of radio frequency signals emitted by atomic spins using the antenna of a magnetic resonance machine during a magnetic resonance imaging scan. Magnetic resonance data is an example of tomographic medical image data.

[0059] A magnetic resonance imaging (MRI) image or MR image is defined herein as a reconstructed two- or three-dimensional visualization of anatomical data contained within magnetic resonance image data. This visualization can be performed using a computer. A three-dimensional magnetic resonance image, or a stack of two-dimensional magnetic resonance images, is an example of three-dimensional tomographic medical image data.

[0060] Computed tomography measurement data is defined as measurements obtained when imaging a subject with a computed tomography system, for example the computed tomography data is an acquired X-ray attenuation profile.

[0061] A computed tomography (CT) image is a reconstructed two- or three-dimensional visualization of the anatomical data contained within the computed tomography measurement data. A three-dimensional computed tomography image, or a stack of two-dimensional computed tomography images, is an example of three-dimensional tomographic medical image data.

[0062] Preferred embodiments of the invention will now be described, by way of example only, with reference to the drawings in which: [Brief description of the drawings]

[0063] [Figure 1] FIG. 1 illustrates an example of a medical system. [Diagram 2] 2 is a flow chart illustrating a method of use of the medical system of FIG. 1. [Diagram 3] FIG. 1 illustrates a further example of a medical system. [Figure 4] FIG. 1 illustrates a further example of a medical system. [Diagram 5] FIG. 2 illustrates a two-dimensional slice of exemplary corrected three-dimensional tomographic medical image data. [Figure 6] FIG. 1 is a diagram illustrating an example of a shape model formed from a plurality of surface meshes. [Figure 7] FIG. 7 illustrates the application of the shape model of FIG. 6 to the magnetic resonance image 122 of FIG. 5. [Figure 8] FIG. 8 is a diagram showing a two-dimensional cross section of the shape model shown in FIGS. 6 and 7. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0064] Like numbered elements in the figures are equivalent elements or perform the same function. An element described earlier is not necessarily described in a later figure if the function is equivalent.

[0065] 1 illustrates an example of a medical system 100. The medical system 100 is shown as including a computer 102 having a computing system 104. The computer 102 is intended to represent one or more computers or computing devices located at one or more locations. Similarly, the computing system 104 may include one or more processing or computing cores and may be located at a single location or distributed.

[0066] The medical system 100 is further shown to include an optional user interface 106. The optional user interface 106 allows, for example, an operator of the medical system 100 to interact with and / or control the medical system 100. The medical system 100 is shown to include an optional network connection 108. The network connection 108 can be used by the medical system 100 to communicate with, for example, another system, for example, the medical system 100 as shown can be remotely located via the Internet and evaluated by various centers where tomographic medical imaging is performed. The medical system 100 is further shown to include a memory 110 accessible to the computing system 104. The memory 110 is intended to represent various types of memory and storage devices accessible to the computing system 104.

[0067] The memory 110 is shown as storing machine executable instructions 120. The machine executable instructions 120 enable the computing system 104 to perform various tasks, such as controlling other components, if present, and performing various image processing and numerical tasks, such as image reconstruction or image processing of tomographic medical imaging data. The memory 110 is further shown as including modified 3D tomographic medical image data 122. The modified 3D tomographic medical image data 122 is face-cleaned in the facial zone, meaning that voxels that are within the subject's facial region are either set to a predefined value or are blurred or filled in such that anatomical details are no longer clear. The memory 110 is further shown as including a segmentation module 124. The segmentation module 124 is executable code that enables fitting a model-based segmentation 126 to the modified 3D tomographic medical image data 122.

[0068] The segmentation module 124, in some examples, is modified so that the fitting parameters in the facial zone do not use anatomical landmarks or anatomical constraints to match the modified 3D tomographic medical image data 122. For example, when using a shape-deformable model, other parts of the modified 3D tomographic medical image data 122 are used and the model-based segmentation 126 is simply fitted to fit the remaining parts of the model. The memory 110 is further shown as including any identification of the location of the facial zone 128. For example, an algorithm can be used to detect the location of voxels where the subject's facial features have been erased. In another example, the identification of the location of the facial zone 128 can be provided as metadata accompanying the modified 3D tomographic medical image data 122.

[0069] For example, if facial features in the face zone are erased by the automatic algorithm, it is possible to attach to the modified three-dimensional tomographic medical image data a description of which voxels have been erased or modified. The memory 110 further includes an identification of the location of the reconstructed region 130, which is the region in which the voxels have been assigned anatomical labels. As previously mentioned, this may be exclusively the face zone, the entire region of interest including the subject's head, or an intermediate value. In some cases, the identification of the location of the reconstructed region 130 is provided by the automatic algorithm or received from the user interface 106. The memory 110 is further shown as including restored three-dimensional tomographic medical image data 132, which has been prepared by replacing voxels in the reconstructed region 130 with anatomical labels assigned to the voxels in the reconstructed region 130. As will be described later, this can be achieved in several different ways.

[0070] FIG. 2 shows a flow chart illustrating a method of operating the medical system of FIG. 1. First, in step 200, modified 3D tomographic medical image data 122 is received. The modified 3D tomographic medical image data 122 describes the head of a subject. The modified 3D tomographic medical image data 122 includes voxels, with facial zones of the medical image data having been facially erased. This means that a portion of the 3D tomographic medical image data 122 describes the facial region of the subject, which is set to a predefined value or blurred so that the facial features are obscured. Next, in step 202, a model-based segmentation 126 is fitted to the modified 3D tomographic medical image data 122. The model-based segmentation 126 comprises a number of surface meshes that define anatomical regions. For example, these surface meshes define the boundaries between different anatomical regions. Thus, the regions between the different surface meshes have a particular anatomical region or tissue type.

[0071] Next, in step 204, each of the voxels of the reconstructed region 130 is labeled with an anatomical label according to the anatomical region defined by the plurality of surface meshes. The reconstructed region includes at least the facial zone. Finally, in step 206, the voxels in the reconstructed region 130 are assigned a contrast value with the anatomical label to provide a restored three-dimensional tomographic medical image data. This can be accomplished in a variety of different ways. In one example, only the voxels in the facial zone 128 are replaced. In another example, all voxels are replaced. In this case, the voxels describing the subject's brain are later pasted into a new image to provide a restored three-dimensional tomographic medical image data. In other cases, the reconstructed region can be an intermediate region including a transition between the facial zone and the remainder of the restored three-dimensional tomographic medical image data.

[0072] Figure 3 shows a further example of a medical system 300. The medical system 300 is similar to the medical system 100 shown in Figure 1, except that the medical system 300 further comprises a magnetic resonance imaging system 302 having a controller 102'.

[0073] The magnetic resonance imaging system 302 comprises a magnet 304. The magnet 304 is a superconducting cylindrical magnet with a bore 306 passing through it. It is also possible to use different types of magnets, 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 iso-face of the magnet; such magnets are used, for example, in combination with charged particle radiation therapy. Open magnets have two magnet parts, one above the other, with a space between them large enough to accommodate the subject, the arrangement of the two part areas being similar to 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.

[0074] Within the bore 306 of the cylindrical magnet 304 is an imaging zone 308 in which there is a magnetic field strong and uniform enough to perform magnetic resonance imaging. Shown within the imaging zone 308 is a field of view 309. Acquired k-space data is typically acquired relative to the field of view 309. The region of interest may be identical to the field of view 309 or may be a sub-volume of the field of view 309. A subject 318 is shown as being supported by a subject support 320 such that at least a portion of the subject 318 is within the imaging zone 308 and the field of view 309.

[0075] Also within the magnet bore 306 are a set of magnetic field gradient coils 310 used for acquiring preliminary k-space data for spatially encoding magnetic spins within the imaging zone 308 of the magnet 304. The magnetic field gradient coils 310 are connected to a magnetic field gradient coil power supply 312. The magnetic field gradient coils 310 are intended to be representative. Typically, the magnetic field gradient coils 310 will include three separate sets of coils for spatially encoding in three orthogonal spatial directions. The magnetic field gradient power supply supplies electrical current to the magnetic field gradient coils 310. The current supplied to the magnetic field gradient coils 310 is controlled as a function of time and may be ramped or pulsed.

[0076] Adjacent to the imaging zone 308 is a radio frequency coil 314 for manipulating the orientation of magnetic spins in the imaging zone 308 and for receiving radio transmissions from spins also in the imaging zone 308. A radio frequency antenna includes multiple coil elements. A radio frequency antenna is also referred to as a channel or an antenna. The radio frequency coil 314 is connected to a radio frequency transceiver 316. The radio frequency coil 314 and the radio frequency transceiver 316 are replaced by separate transmit and receive coils and separate transmitters and receivers. It is understood that the radio frequency coil 314 and the radio frequency transceiver 316 are representative. The radio frequency coil 314 is also intended to represent a dedicated transmit antenna and a dedicated receive antenna. Similarly, the transceiver 316 also represents a separate transmitter and receiver. The radio frequency coil 314 also has multiple receive / transmit elements, and the radio frequency transceiver 316 has multiple receive / transmit channels.

[0077] The transceiver 316 and the gradient controller 312 are shown as being connected to the hardware interface 106 of the computer system 102 .

[0078] The controller 102' is similar to the computer 102. It comprises a computing system 104' and a network interface 108'. In addition, there is a hardware interface 340 that allows the computing system 104' to control the operation and functions of the magnetic resonance imaging system 302. The controller 102' further comprises a memory 110', which is also representative of various types of memory and storage that the computing system 104' can access. In some instances, it is possible to combine the functions of the controller 102' and the computer 102. However, the computer 102 is often a separate system that accesses the modified three-dimensional tomographic medical image data 122 from a database or storage system.

[0079] The memory 110' is shown as including machine-executable instructions 120', including instructions that enable the computing system 104' to control the magnetic resonance imaging system 302 using the hardware interface 340, including sending commands for control and receiving data acquired by the magnetic resonance imaging system 302.

[0080] The memory 110' is further shown to include pulse sequence commands 330 that are used by the computing system 104' to control the magnetic resonance imaging system 302 to acquire k-space data 332 describing the field of view 309. The memory 110' is further shown to include three-dimensional magnetic resonance image data 334 that has been reconstructed from the k-space data 332. The k-space data 332 has been acquired by controlling the magnetic resonance imaging system 302 with the pulse sequence commands 330. The memory 110' is further shown to include modified three-dimensional tomographic medical image data 122. In this example, the modified three-dimensional tomographic medical image data 122 was constructed from the three-dimensional magnetic resonance image data 334 by erasing voxels of the facial zone from the face. This can be performed manually by an operator or by an automated algorithm that can be programmed into the machine executable instructions 120', for example.

[0081] A network connection 338 between the network interfaces 108 and 108' allows the computer 102 and the controller 102' to exchange data. In this case, the corrected three-dimensional tomographic medical image data 122 was sent from the controller 102' to the computer 102. The computer 102 shown in FIG. 3 is similar to the computer 102 shown in FIG. 1, except that the memory 110 is further shown to include a reconstruction neural network 336. The reconstruction neural network 336 takes as input the mapping between the anatomical region and at least the reconstructed region 134 and the corrected three-dimensional tomographic medical image data 122. In response, the reconstruction neural network 336 outputs the generated three-dimensional magnetic resonance image data 342.

[0082] The generated three-dimensional magnetic resonance image data 342 is essentially a composite image generated by the reconstruction neural network 336. The voxels of the generated three-dimensional magnetic resonance image data 342 and the corrected three-dimensional tomographic medical image data 122 can be used to construct the restored three-dimensional tomographic medical image data 132. For example, the voxels in the reconstructed region 130 can be pasted into the corrected three-dimensional tomographic medical image data 122. Alternatively, voxels can be pasted into the corrected three-dimensional tomographic medical image data 122 only from the generated three-dimensional magnetic resonance image data 342. There are various different ways in which the corrected three-dimensional tomographic medical image data 122 can be reconstructed. In another example, the entire generated three-dimensional magnetic resonance image data 342 is used, and only voxels from the corrected three-dimensional tomographic medical image data 122 that encompass the brain volume are pasted into the generated three-dimensional magnetic resonance image data 342 to create the restored three-dimensional tomographic medical image data 132. In this example, anatomical landmarks etc. from the generated three-dimensional magnetic resonance image data 342 are used, and then data describing the subject's brain from the original images is used.

[0083] The memory 110 is further shown as including a mapping of at least the anatomical regions of the reconstructed region 134. This mapping is an assignment of anatomical regions to the various voxels in the reconstructed region 130. The model-based segmentation 126 divides the corrected three-dimensional tomographic medical image data 122 into different anatomical regions. It is then very simple to assign a value to each of these.

[0084] Figure 4 shows a further example of a medical system 400. The medical system 400 of Figure 4 is similar to the medical system 300 of Figure 3, except that instead of comprising a magnetic resonance imaging system 302, the medical system 400 of Figure 4 comprises a computed tomography system 402 rather.

[0085] The CT system 402 includes a rotating gantry 404. The gantry 404 rotates about an axis of rotation 406. A subject 318 lies on a subject support 320. Inside the gantry 404 is an x-ray tube 410.

[0086] X-ray tube 410 produces x-rays 414 that pass through collimator 411 and then through subject 318 before being received by detector 412. Within the area of ​​box 416 is an imaging zone where a CT or computed tomography image of subject 418 may be acquired. CT system 402 is shown as being controlled by controller 102'. Hardware interface 304 allows processor 106 to exchange messages and control CT system 502.

[0087] In this example, the contents of the memory 110' of the controller 102' are varied. In this example, there are machine executable instructions 120" that further provide image processing tasks, which also enable the computing system 104' to control the computed tomography system 402 via the hardware interface 340. The memory 110' further includes CT system control commands 430. These are in particular commands that the computing system 104' uses to control the computed tomography system 402 to acquire computed tomography measurement data 432. The machine executable instructions 120" also enable the computing system 104' to reconstruct three-dimensional computed tomography data 434 from the computed tomography measurement data 432. And again, the machine executable instructions 120" cause the computing system 104' to construct modified three-dimensional tomography medical image data 122 from the three-dimensional computed tomography data 434. The controller 102' transmits the modified three-dimensional tomography medical image data 122 to the computer 102 via the network connection 338.

[0088] In the example of FIG. 3, the reconstruction neural network 336 was used. In this example, a lookup table 440 is used instead. The lookup table 440 obtains a mapping of anatomical regions in at least the reconstructed region 134, and then assigns an intensity or contrast to each of these voxels using the lookup table 440. In some examples, the values ​​in the lookup table 440 are adjusted by looking up regions in the corrected three-dimensional tomographic medical image data 122 to match the contrast. After the voxels in the reconstructed region 130 are replaced, the corrected three-dimensional tomographic medical image data 122 is converted to a restored three-dimensional tomographic medical image data 132.

[0089] In some clinical applications, facial parts are necessary for subsequent algorithmic processing of images, e.g. for volume conductor generation for electrophysiological applications like EEG / MEG or TMS / TDCS / TACS. Furthermore, segmentation algorithms may also benefit from facial context information.

[0090] In current applications, the facial region is simply removed, resulting in unrealistic shapes. This can cause serious problems, e.g., regarding the generation of volume conductors, due to modeling inaccuracies and the important role of hard tissues such as bones. This is especially relevant for Finite Element Modeling (FEM), which is usually more accurate compared to Boundary Element Modeling (BEM).

[0091] 5 shows a single slice of the corrected 3D tomographic medical image data 122. In this example, the corrected 3D tomographic medical image data 122 is a magnetic resonance image. It can be clearly seen that there is a facial zone 128 where the facial features have been removed by being set to the same value as the background.

[0092] In an example, a tool is provided to substitute a patient's face that has been removed from an MRI or other tomography scan, which may be accomplished by a series of processing steps: a) optionally removing facial features with a classical de-identification algorithm, b) applying model-based segmentation (MBS) to the new image to generate a smooth representation of individual anatomical structures and a generic face, c) generating an image from a surface mesh of the MBS, and c) integrating generic facial features into the image from which the patient's face has been removed.

[0093] A triangular surface model is constructed that includes anatomical structures such as skin, skull, air, and brain hemispheres. This surface model is trained to adapt to unseen MR images.

[0094] 6 shows an example of a geometric model 600 in which several surface meshes are visible. The region labeled 602 is a surface mesh that defines the boundary of the skin. The region labeled 606 represents the surface that defines the airway 606 or air cavity. The circular region labeled 608 defines the boundary of the subject's eyes 608.

[0095] During adaptation, each mesh triangle searches for a target point in the image space along its normal. The target point is a characteristic image feature (typical gray value, edge) that was previously learned during model construction. The target point found by a triangle is called E ext The so-called extrinsic energy, called extrinsic energy, is included in the estimation of the extrinsic energy. Triangles far from the image features do not seek the target point and therefore do not contribute to the extrinsic energy. Thus, the extrinsic energy is driven by the parts of the image that contain the image features, while the face-erased regions are ignored in the calculation of the extrinsic energy. This can be expressed as

number

number

[0096] In contrast to external energy, internal energy E int does not take into account any image information or image features. It is concerned with a) penalizing the deviation between the current mesh state and the mesh from the personalized 3D model, and b) ensuring shape consistency. This is given by the following equation:

number

[0097] The final deformation is achieved by minimizing the sum of the internal and external energies. E:=αE int +(1-α)E ext α is a parameter to balance the influence of each energy. Different parts of the mesh can contribute by having different internal energy weights or stiffness. In our case, the face region will have a significantly larger value.

[0098] Figure 7 shows the fit of the shape model of Figure 6 to the magnetic resonance image 122 of Figure 5. Note that this is for illustrative purposes only and the model is not optimized for this application, so the mesh boundaries are not smooth. Figure 7 further illustrates the magnetic resonance image shown in Figure 5. In this case, a model-based segmentation 126 has been fitted to the magnetic resonance image 122. It can be seen that the various elements of the model-based segmentation 126 extend into the facial zone 128. The surface mesh representing the skin 602 is shown as being on the outside and is also shown in the right-hand view of the magnetic resonance image 122.

[0099] After model-based segmentation, the surface mesh needs to be transformed into voxel space. This is accomplished by labeling each voxel within a particular region as a member of this region. Figure 8 shows an example. Finally, machine learning algorithms for segmentation, image classification, or FEM algorithms can utilize the segmentation map that incorporates a generic face.

[0100] 8 shows an example of a segmentation map 134 (or mapping of anatomical regions) incorporating several surface meshes 700 that define distinct anatomical regions 802. This can be used together with a look-up table to restore the corrected 3D tomographic medical image data 122 or also used as input to the reconstruction neural network 336. This is realized for example as a so-called generative adversarial network (GAN), where an artificial neural network learns to map images with faces erased onto images with faces.

[0101] Also displayed in the segmentation 134 is a contour representing the boundary of the brain volume 802. In some examples, voxels from the brain volume 802 of the corrected 3D tomographic medical image data are pasted or copied into the corresponding brain volume 802 of the restored 3D tomographic medical image data 132. This ensures that during clinical examination, the voxels in the brain volume 802 are of the subject and not synthetically generated.

[0102] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description is to be considered illustrative or exemplary and not restrictive, and the invention is not limited to the disclosed embodiments.

[0103] Other variations to the disclosed embodiments can be understood and realized by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does 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 another hardware, but may also be distributed in another form, such as via the Internet or another wired or wireless telecommunication system. Any reference signs in the claims should not be interpreted as limiting the scope. [Explanation of symbols]

[0104] 100 Medical Systems 102 Computer 104 Computing Systems 106 Optional User Interface 108 Network Connection 110 Memory 120 Machine Executable Instructions 122 Corrected 3D Tomographic Medical Image Data 124 Segmentation Module 126 Model-Based Segmentation 128 facial zone location identifiers 130 Identification of the location of the reconstructed region 132 Reconstructed 3D Tomographic Medical Image Data 134 Mapping of anatomical regions at least in the reconstructed area 200 receiving corrected three-dimensional tomographic medical image data including voxels describing the subject's head; 202 A model-based segmentation is applied to the corrected 3D tomographic medical image data, where the model-based segmentation models the entire head. 204 For each voxel of a reconstructed region of the corrected 3D tomographic medical image data, an anatomical label is assigned according to an anatomical region defined by the plurality of surface meshes. 206 Assign contrast values ​​to voxels in the reconstructed regions using anatomical labels to provide restored 3D tomographic medical image data 300 Medical Systems 302 Magnetic Resonance Imaging System 304 Magnet 306 Magnet Bore 308 Imaging Zone 309 Field of view 310 Magnetic field gradient coil 312 Magnetic field gradient coil power supply 314 Radio Frequency Coil 316 Transceiver 318 Target 320 Subject Support 330 Pulse Sequence Commands 332 k-space data 334 Three-dimensional magnetic resonance imaging data 336 Reconstruction Neural Networks 338 Network Connection 340 Hardware Interface 342 Generated 3D magnetic resonance image data 400 Medical Systems 402 CT System 404 Gantry 406 Rotational Axis 410 X-ray tube 411 Collimator 412 Detector 414 X-ray 416 Imaging Zone 420 Subject Support Actuator 430 CT system control command 432 Computed tomography measurement data 434 Three-dimensional computed tomography data 440 Lookup Table 600 Shape Model 602 Skin 604 Skull 606 Airway 608 eyes 700 Surface Mesh 800 distinct anatomical regions 802 Brain Volume

Claims

1. a memory storing machine-executable instructions; a computing system, wherein execution of the machine-executable instructions causes the computing system to: receiving modified three-dimensional tomographic medical image data including voxels describing a subject's head having a facial obliteration within a facial zone; fitting a model-based segmentation to the corrected 3D tomographic medical image data, the model-based segmentation modeling an entire head, the model-based segmentation comprising a plurality of surface meshes defining anatomical regions; assigning an anatomical label to each of the voxels of a reconstructed region of the corrected 3D tomographic medical image data according to an anatomical region defined by the plurality of surface meshes, the reconstructed region including at least the facial zone; and assigning contrast values ​​to voxels of the reconstructed region using the anatomical labels to provide restored three-dimensional tomographic medical image data, wherein assigning contrast values ​​includes assigning intensities; and The medical system that makes it possible for

2. 2. The medical system of claim 1, wherein execution of the machine-executable instructions further causes the computing system to use the restored three-dimensional tomographic medical image data in one of a segmentation algorithm, an electroencephalogram algorithm, a magnetoencephalogram algorithm, a transcranial magnetic stimulation planning algorithm, a transcranial direct current stimulation planning algorithm, and a transcranial alternating current stimulation planning algorithm.

3. 3. The medical system of claim 2, wherein execution of the machine-executable instructions further causes the computing system to use the model-based segmentation as input to any one of the electroencephalogram algorithm, the magnetoencephalogram algorithm, the transcranial magnetic stimulation planning algorithm, the transcranial direct current stimulation planning algorithm, and the transcranial alternating current stimulation planning algorithm for electromagnetic computation within the head of the subject.

4. 4. The medical system of claim 1, wherein execution of the machine-executable instructions causes the computing system to assign the contrast values ​​to voxels in the reconstructed region using a lookup table with an entry for each anatomical label.

5. 5. The medical system of claim 4, wherein execution of the machine-executable instructions causes the computing system to normalize the lookup table with regions of the corrected 3D tomographic medical image other than the facial zone before assigning the contrast values ​​to voxels in the reconstructed region.

6. The medical system of claim 5 , wherein execution of the machine-executable instructions further causes the computing system to perform smoothing and / or contrast matching at boundaries of the reconstructed region.

7. 4. The medical system of claim 1, wherein the memory further comprises a reconstruction neural network that assigns the contrast value to the voxels of the reconstructed region in response to receiving the corrected three-dimensional tomographic medical image data and the anatomical label of each voxel of the reconstructed region.

8. 2. The medical system of claim 1, wherein the model-based segmentation includes a brain volume, and execution of the machine-executable instructions causes the computing system to ensure that the brain volume of the restored three-dimensional tomographic medical image data is identical to the brain volume of the corrected three-dimensional tomographic medical image data.

9. The medical system of claim 1 , wherein the corrected three-dimensional tomographic medical image data is a three-dimensional data set or a stack of two-dimensional slices.

10. The medical system of claim 1 , wherein the modified three-dimensional tomographic medical image data is one of magnetic resonance image data and computed tomography image data.

11. The medical system of claim 1 , wherein the model-based segmentation is one of deformable shape model segmentation, active contour model, and locally deformable statistical shape model segmentation.

12. Execution of the machine-executable instructions further causes the computing system to: receiving acquired three-dimensional tomographic medical image data; receiving the modified three-dimensional tomographic medical image data in response to inputting the acquired three-dimensional tomographic medical image data into a face removal algorithm; The medical system according to claim 1 ,

13. 10. The medical system of claim 1, further comprising a tomographic medical imaging system, and execution of the machine-executable instructions further causes the computing system to control the tomographic medical imaging system to obtain the acquired three-dimensional tomographic medical image data.

14. The medical system of claim 1 , wherein voxels within the face zone are facially eliminated by deletion or blurring.

15. The medical system of claim 14 , wherein the removal of voxels from the facial zone comprises setting a voxel to a predetermined value or removing a value assignment.

16. The medical system of claim 1 , wherein the facial zone comprises at least a portion of the facial skeleton.

17. The medical system of claim 1 , wherein the reconstructed region is within the anatomical region.

18. 2. The medical system of claim 1, wherein each of the voxels of the reconstructed region of the corrected three-dimensional tomographic medical image data is assigned a corresponding anatomical label according to the anatomical region defined by the plurality of surface meshes.

19. receiving modified three-dimensional tomographic medical image data including voxels describing a subject's head with facial obliteration within a facial zone; fitting a model-based segmentation to the corrected 3D tomographic medical image data, the model-based segmentation models an entire head, the model-based segmentation comprising a plurality of surface meshes defining anatomical regions; assigning an anatomical label to each of the voxels of a reconstructed region of the corrected 3D tomographic medical image data according to an anatomical region defined by the plurality of surface meshes, the reconstructed region including at least the facial zone; assigning contrast values ​​to voxels of the reconstructed region using the anatomical labels to provide restored three-dimensional tomographic medical image data, wherein assigning contrast values ​​includes assigning intensities; 1. A method of medical imaging comprising:

20. 1. A computer program comprising machine-executable instructions for execution by a computing system, wherein execution of the machine-executable instructions causes the computing system to: receiving modified three-dimensional tomographic medical image data including voxels describing a subject's head having a facial obliteration within a facial zone; fitting a model-based segmentation to the corrected 3D tomographic medical image data, the model-based segmentation modeling an entire head, the model-based segmentation comprising a plurality of surface meshes defining anatomical regions; assigning an anatomical label to each of the voxels of a reconstructed region of the corrected 3D tomographic medical image data according to an anatomical region defined by the plurality of surface meshes, the reconstructed region including at least the facial zone; and assigning contrast values ​​to voxels of the reconstructed region using the anatomical labels to provide restored three-dimensional tomographic medical image data, wherein assigning contrast values ​​includes assigning intensities; and A computer program that performs the following:

Images