Apparatus for computed tomography x-ray data acquired at high relative pitch

Abstract

The invention relates to an apparatus (10) for correcting computed tomography ("CT") X-ray data acquired at high relative pitch, the apparatus comprising: an input unit (20), a processing unit (30), and an output unit (40). The input unit is configured to provide the processing unit with CT X-ray data of a body part of a person acquired at high relative pitch. The processing unit is configured to determine CT slice reconstruction data of the body part of the person without or with reduced high relative pitch operation reconstruction artifacts using a machine learning algorithm. The machine learning algorithm is trained based on CT slice reconstruction data, and wherein the CT slice reconstruction data comprises first CT slice reconstruction data with high relative pitch reconstruction artifacts, and comprises second CT slice reconstruction data without or with less or less severe high relative pitch reconstruction artifacts. The output unit is configured to output the CT slice reconstruction data of the body part of the person.

Description

TECHNICAL FIELD

[0001] The present application relates to an apparatus for correcting CT (Computed Tomography) X-ray data acquired at high relative pitch, an imaging system for correcting CT X-ray data acquired at high relative pitch, a method and a computer program element and a computer readable medium. BACKGROUND

[0002] In helical CT, the pitch is the speed of the patient table in the z-direction per unit of gantry rotation, and it determines the speed at which a certain volume can be scanned. The relative pitch is the ratio between this absolute pitch value and the size of the volume covered in the z-direction by a given source / detector geometry at the center of rotation. In typical cases, the relative pitch is usually between 0.2 and 1.2, depending on the protocol.

[0003] For certain clinical protocols that are forced to cover large volumes in very short time periods, high pitches must be used. Examples of such applications are single beat cardiac scans and peripheral bolus runoff protocols. The problem that arises for high relative pitches is that at least 180° of X-ray does not see certain outer volume regions in the x / y plane, which is a necessary condition for slice reconstruction. This deficiency often introduces artifacts in the image.

[0004] Dual source systems have been developed to overcome these limitations for cardiac imaging. The shifted dual tube / detector geometry allows for a pitch that is about twice the pitch in the case of single source CT, but this solution doubles the hardware cost and has other drawbacks, e.g. sensitivity to X-ray backscatter.

[0005] Alternatively, the detector can be combined with a larger z coverage, but this again increases the total cost of the system.

[0006] EP 3467766 discloses the use of machine learning to correct CT images.

[0007] There is a need to address these problems. SUMMARY

[0008] It would be advantageous to have an improved means to correct computed tomography X-ray data acquired at high relative pitch. It is to be noted that the various aspects and examples of the application described below also apply to an apparatus for correcting computed tomography X-ray data acquired at high relative pitch, an imaging system for correcting computed tomography X-ray data acquired at high relative pitch, a method and a computer program element and a computer readable medium.

[0009] In a first aspect, there is provided an apparatus for correcting CT X-ray data acquired at high relative pitch, the apparatus comprising:

[0010] an input unit;

[0011] a processing unit; and

[0012] an output unit;

[0013] The input unit is configured to provide the processing unit with CT X-ray data of a body part of a person acquired at high relative pitch. The processing unit is configured to determine CT slice reconstruction data of the body part of the person without high relative pitch operation reconstruction artifacts or with reduced high relative pitch operation reconstruction artifacts using a machine learning algorithm. The machine learning algorithm is trained based on CT slice reconstruction data, and wherein the CT slice reconstruction data comprises first CT slice reconstruction data with high relative pitch reconstruction artifacts, and comprises second CT slice reconstruction data without high relative pitch reconstruction artifacts or with less or less severe high relative pitch reconstruction artifacts. The output unit is configured to output the CT slice reconstruction data of the body part of the person.

[0014] In an example, the CT X-ray data for the second CT slice reconstruction data is acquired at low relative pitch.

[0015] In an example, the CT X-ray data for the second CT slice reconstruction data comprises full angular coverage data.

[0016] In an example, the CT X-ray data for the second CT slice reconstruction data comprises CT X-ray data of the body part of one or more test persons.

[0017] In an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data for the second CT slice reconstruction data that has been manipulated.

[0018] In an example, the manipulation comprises simulating the CT X-ray data from the second CT slice reconstruction data in order to provide CT X-ray data for the first CT slice reconstruction data that would have been effectively acquired at a higher relative pitch than the relative pitch used to acquire the CT X-ray data for the second CT slice reconstruction data.

[0019] In an example, the manipulating comprises simulating the CT X-ray data for the second CT slice reconstruction data so as to provide CT X-ray data for the first CT slice reconstruction data that is effectively acquired by a resized detector, the resized detector being resized for a detector used to acquire the CT X-ray data for the second CT slice reconstruction data (i.e. the CT X-ray data for the first CT slice reconstruction data is simulated to be acquired by a resized detector, the resized detector being resized for a detector used to acquire the CT X-ray data for the second CT slice reconstruction data).

[0020] In an example, the simulating comprises reducing a size of the detector in a direction parallel to an axis of rotation of an X-ray imaging system used to acquire the CT X-ray data for the second CT slice reconstruction data.

[0021] In an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data of one or more test persons.

[0022] In an example, the training of the machine learning algorithm comprises registering at least some of the second CT slice reconstruction data with the first CT slice reconstruction data reconstructed from the CT X-ray data of one or more test persons.

[0023] In an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data of the body part of the one or more test persons.

[0024] In a second aspect, there is provided an imaging system comprising:

[0025] an X-ray source;

[0026] an X-ray detector; and

[0027] an apparatus for correcting computed tomography, “CT”, X-ray data acquired at high relative pitch according to the first aspect.

[0028] The X-ray source and the X-ray detector are configured to rotate around a body part of a person and to acquire CT X-ray data at high relative pitch. The apparatus is configured to output CT slice reconstruction data of the body part of the person.

[0029] In a third aspect, there is provided a method for correcting computed tomography, “CT”, X-ray data acquired at high relative pitch, the method comprising:

[0030] a) providing a processing unit with CT X-ray data of a body part of a person acquired at high relative pitch;

[0031] b) determining, with the processing unit, CT slice reconstruction data of the body part of the person without or with reduced high relative pitch operation reconstruction artifacts using a machine learning algorithm; wherein the machine learning algorithm is trained based on CT slice reconstruction data, and wherein the CT slice reconstruction data comprises first CT slice reconstruction data with high relative pitch reconstruction artifacts, and second CT slice reconstruction data without or with less or less severe high relative pitch reconstruction artifacts; and

[0032] c) outputting, by an output unit, the CT slice reconstruction data of the body part of the person.

[0033] According to another aspect, there is provided a computer program element controlling one or more of the aforementioned apparatuses or systems, which, when being executed by a processing unit, is adapted to perform one or more of the aforementioned methods.

[0034] According to another aspect, there is provided a computer readable medium having the aforementioned computer element stored therein.

[0035] The computer program element might, for example, be a software program but could also be a FPGA, a PLD or any other appropriate digital means.

[0036] Advantageously, the benefits provided by any of the above aspects apply equally to all the other aspects, and vice versa.

[0037] The above aspects and examples will become apparent and elucidated with reference to the embodiments described hereinafter for description of which reference is made to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0038] In the following exemplary embodiments will be described with reference to the accompanying drawings:

[0039] Figure 1 A schematic setup of an example of an apparatus for correcting computed tomography X-ray data acquired at high relative pitch is shown;

[0040] Figure 2 A schematic setup of an example of an imaging system is shown;

[0041] Figure 3 A method for correcting computed tomography X-ray data acquired at high relative pitch is shown;

[0042] Figure 4 An example of a setup for training a neural network is shown;

[0043] Figure 5 An example of using a trained neural network to generate corrected images is shown;

[0044] Figure 6 An example of two slices corrected using a trained neural network is shown;

[0045] Figure 7 A detailed example of the imaging system is shown; and

[0046] Figure 8 A schematic setup showing a person moving via a rotating CT X-ray system is illustrated. Detailed Implementation

[0047] Figure 1 An example of an apparatus 10 for correcting CT X-ray data acquired under high relative pitch is shown. Apparatus 10 includes an input unit 20, a processing unit 30, and an output unit 40. The input unit is configured to provide the processing unit with CT X-ray data of a person's body portion acquired under high relative pitch. The processing unit is configured to use a machine learning algorithm to determine CT slice reconstruction data of the person's body portion that is free of high relative pitch reconstruction artifacts or has reduced high relative pitch reconstruction artifacts. The machine learning algorithm is trained based on the CT slice reconstruction data. The CT slice reconstruction data includes first CT slice reconstruction data with high relative pitch reconstruction artifacts and second CT slice reconstruction data that is free of high relative pitch reconstruction artifacts or has fewer or less severe high relative pitch reconstruction artifacts. The output unit is configured to output the CT slice reconstruction data of the person's body portion.

[0048] In the example, the machine learning algorithm is a neural network.

[0049] In the example, the neural network is trained in a supervised manner, so that it knows the desired output.

[0050] In the example, the neural network is trained in an unsupervised manner, so that the expected output is unknown.

[0051] In the example, the neural network is trained in an unsupervised manner using a generative adversarial neural network (GAN) method. The relative pitch involves the following: the distance the patient stage of the X-ray system travels in the axial direction from one rotation to the next, divided by the projection of the detector backward toward the X-ray source at its axial position.

[0052] High relative pitch here refers to relative pitches greater than 1 and usually less than 2, but it also applies to relative pitches greater than 2.

[0053] Low relative pitch involves a relative pitch equal to or less than 1.

[0054] According to the example, the CT X-ray data for the reconstruction of the second CT slice was acquired at a low relative pitch.

[0055] According to the example, the CT X-ray data for the reconstruction data of the second CT slice includes full angular coverage data.

[0056] In the example, complete angle data includes data collected at rotation angles of at least 180 degrees.

[0057] According to the example, the CT X-ray data for the reconstruction data of the second CT slice includes CT X-ray data of one or more body parts of the test subject.

[0058] According to the example, the CT X-ray data for the reconstruction data of the first CT slice includes the CT X-ray data for the reconstruction data of the second CT slice that has been manipulated.

[0059] In the example, the CT X-ray data for the reconstruction of the first CT slice cannot be reconstructed without artifacts.

[0060] According to the example, the manipulation includes simulating CT X-ray data based on second CT slice reconstruction data to provide CT X-ray data for the first CT slice reconstruction data that is effectively acquired at a higher relative pitch than the relative pitch used to acquire the CT X-ray data for the second CT slice reconstruction data.

[0061] According to the example, the manipulation includes simulating CT X-ray data for the reconstruction data of the second CT slice in order to provide CT X-ray data for the reconstruction data of the first CT slice: the CT X-ray data for the reconstruction data of the first CT slice is effectively acquired by a resized detector, the resized detector being resized for a detector used to acquire CT X-ray data for the reconstruction data of the second CT slice (i.e., the CT X-ray data for the reconstruction data of the first CT slice is simulated as being acquired by a resized detector, the resized detector being resized for a detector used to acquire CT X-ray data for the reconstruction data of the second CT slice).

[0062] According to the example, the simulation involves reducing the size of the detector in a direction parallel to the rotation axis of the X-ray imaging system used to acquire CT X-ray data for reconstructing data from a second CT slice.

[0063] Thus, the size of the detector is reduced in a direction, typically the row direction.

[0064] In an example, the simulation comprises simulating data acquisition by a detector reduced in the row direction to half of the original size.

[0065] According to an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data of one or more test persons.

[0066] In an example, the CT X-ray data for the first CT slice reconstruction data fails to enable a reconstruction of the first CT slice reconstruction data without artifacts.

[0067] In an example, the CT X-ray data for the first CT slice reconstruction data is acquired at a high relative pitch.

[0068] In an example, the CT X-ray data for the first CT slice reconstruction data comprises incomplete angular coverage data.

[0069] In an example, the incomplete angular data comprises data acquired at a rotation angle of less than 180 degrees.

[0070] According to an example, the training of the machine learning algorithm comprises registering at least some of the second CT slice reconstruction data with the first CT slice reconstruction data reconstructed from the CT X-ray data of the one or more test persons.

[0071] According to an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data of a body part of one or more test persons.

[0072] Figure 2 An example of an imaging system 50 is shown. The imaging system 50 comprises an X-ray source 60, an X-ray detector 70 and a device 10 for correcting CT X-ray data acquired at a high relative pitch as described with respect to Figure 1 The X-ray source and the X-ray detector are configured to rotate around a body part of a person and to acquire CT X-ray data at a high relative pitch. The device is configured to output CT slice reconstruction data of the body part of the person.

[0073] Thus, in a clinical setting, a patient’s organ of interest (e.g. the heart) can be positioned in or near the rotation axis of the system. This can minimize the amount of artifacts in the clinically relevant part of the image. Thus, even at high pitch, this target volume typically only requires a small amount of correction. However, those parts of the image volume that are (to some extent) off-axis from the rotation axis suffer from artifacts, which can now be corrected by the trained machine learning algorithm.

[0074] Figure 3 A method 200 for correcting CT X-ray data acquired at high relative pitch is shown. The method comprises:

[0075] In a providing step 210 (also referred to as step a), CT X-ray data of a body part of a person acquired at high relative pitch is provided to a processing unit;

[0076] In a determining step 220 (also referred to as step b), CT slice reconstruction data of the body part of the person without high relative pitch operation reconstruction artifacts or with reduced high relative pitch operation reconstruction artifacts is determined with the processing unit using a machine learning algorithm; wherein the machine learning algorithm is trained based on CT slice reconstruction data, and wherein the CT slice reconstruction data comprises first CT slice reconstruction data with high relative pitch reconstruction artifacts, and comprises second CT slice reconstruction data without high relative pitch reconstruction artifacts or with less or less severe high relative pitch reconstruction artifacts; and

[0077] In an outputting step 230 (also referred to as step c), the CT slice reconstruction data of the body part of the person is output by an output unit.

[0078] In an example, the machine learning algorithm is a neural network.

[0079] In an example, the neural network is trained in a supervised manner, such that the desired output is known.

[0080] In an example, the neural network is trained in an unsupervised manner, such that the desired output is not known.

[0081] In an example, the neural network is trained in an unsupervised manner by using a generative adversarial neural network (GAN) approach.

[0082] In an example, the CT X-ray data for the second CT slice reconstruction data enables a reconstruction of the second CT slice reconstruction data without artifacts.

[0083] In an example, the CT X-ray data for the second CT slice reconstruction data is acquired at low relative pitch.

[0084] In an example, the CT X-ray data for the second CT slice reconstruction data comprises full angular coverage data.

[0085] In an example, the full angular data comprises data acquired at a rotation angle of at least 180 degrees.

[0086] In an example, the CT X-ray data for the second CT slice reconstruction data comprises CT X-ray data of body parts of one or more test persons.

[0087] In an example, the method comprises manipulating the CT X-ray data for the second CT slice reconstruction data to generate CT X-ray data for the first CT slice reconstruction data.

[0088] In an example, the CT X-ray data for the first CT slice reconstruction data is not capable of enabling a reconstruction of the first CT slice reconstruction data without artifacts.

[0089] In an example, the manipulating comprises simulating the CT X-ray data for the second CT slice reconstruction data in order to provide CT X-ray data for the first CT slice reconstruction data that is effectively acquired at a higher relative pitch than a relative pitch used for acquiring the CT X-ray data for the second CT slice reconstruction data.

[0090] In an example, the manipulating comprises simulating the CT X-ray data for the second CT slice reconstruction data in order to provide CT X-ray data for the first CT slice reconstruction data that is effectively acquired by a resized detector that is resized with respect to a detector used for acquiring the CT X-ray data for the second CT slice reconstruction data (i.e. the CT X-ray data for the first CT slice reconstruction data is simulated as being acquired by a resized detector that is resized with respect to a detector used for acquiring the CT X-ray data for the second CT slice reconstruction data).

[0091] In an example, the simulating comprises reducing a size of the detector in a row direction.

[0092] In an example, the simulating comprises simulating a data acquisition by a detector that is reduced in a row direction to half of an original size.

[0093] In an example, the CT X-ray data for the first CT slice reconstruction data comprises CT X-ray data of one or more test persons.

[0094] In an example, the CT X-ray data for the first CT slice reconstruction data is not capable of enabling a reconstruction of the first CT slice reconstruction data without artifacts.

[0095] In an example, the CT X-ray data for the first CT slice reconstruction data is acquired at a high relative pitch.

[0096] In an example, the CT X-ray data for the first CT slice reconstruction data includes incomplete angular coverage data.

[0097] In an example, the incomplete angular data includes data acquired at less than 180 degrees of rotation angle.

[0098] In an example, the training of the machine learning algorithm includes registering at least some of the second CT slice reconstruction data with the first CT slice reconstruction data reconstructed from the CT X-ray data of the one or more test subjects.

[0099] In an example, the CT X-ray data for the first CT slice reconstruction data includes CT X-ray data of a body part of the one or more test subjects.

[0100] Thus, there is provided a technique in which CT data is acquired by a single-source CT system using an overpitch trajectory, and then a trained machine learning algorithm is utilized to correct for artifacts that arise.

[0101] Reference is now made to Figures 4-8 to describe in more detail an apparatus for correcting computed tomography “CT” X-ray data acquired at high relative pitch and a method for correcting computed tomography “CT” X-ray data acquired at high relative pitch. In the following detailed discussion, a convolutional neural network (CNN) is referenced as the trained machine learning algorithm, but other machine learning algorithms can also be utilized.

[0102] Thus, a deep learning approach is used to remove artifacts generated due to system overpitch from the reconstruction results. In this way, even if the data constraints for image reconstruction are violated, an image without artifacts or with fewer or less severe artifacts is generated.

[0103] The neural network is trained to predict artifacts that arise when CT data is acquired from low relative pitch (e.g., 1.0) to high relative pitch (e.g., 2.0) and correct for these artifacts. In particular embodiments, the artifact-free images (or images with fewer or less severe artifacts) used for training are generated from complete data acquisition at low pitch, while the images with artifacts used for training are images acquired at high pitch. In particular, the inventors have found that using artifact-free data can effectively generate simulated data with artifacts, thereby ensuring complete image registration and avoiding the need for two or more scans of the same subject. This is accomplished by a simulation that involves resizing the detectors used in the original measurements to half of the original size in the row direction.

[0104] In Figure 4The training setup is shown in. At "A1" a schematic representation of the acquisition geometry for full angular coverage using regular pitch is shown. At "A2" a schematic representation of the acquisition geometry for partial angular coverage using high pitch is shown. At "B1" a representation of an example slice of full angular coverage is shown and at "B2" a grey scale encoded angular range acquired in an example slice in the case of partial angular coverage is shown. At "C1" a reconstruction result generated from the data acquired for regular pitch is shown, where the image is free of artifacts. At "C2" a reconstruction result generated from the data acquired for high pitch is shown, where the image has artifacts (indicated by white arrows). Also shown is a difference image between the two slices, which shows the artifacts in the corresponding slice. Then, in this example, the reconstructed image with artifacts and the difference image are used together for the training of a convolutional neural network. Alternatively (not shown in Figure 4 ), the actual image with artifacts and the actual image without artifacts can be used as input to the neural network. The skilled person will appreciate that these data can be used in a number of different ways to train the neural network. In Figure 4 , with respect to the images acquired with low pitch and high pitch, the amount of data available for reconstruction is indicated in the angular coverage images (B1 and B2). For the low pitch case, the angular coverage is consistently completed, i.e. a high value in the circular field of view. For the high pitch case, there is a region left in the image where the amount of data in the parallel beam (wedge) geometry drops below the required 180° of data.

[0105] In Figure 5 , the network inference is depicted, where the prediction of the network is used to produce a corrected image. At "A2" a schematic representation of the acquisition geometry for partial angular coverage using high pitch is shown. At "B2" a grey scale encoded angular range acquired in an example slice in the case of partial angular coverage is shown. At "C2" a reconstruction result generated from the data acquired for high pitch is shown, where the image has artifacts. Now, this data is provided to the trained neural network, which generates a difference image, which can then be utilized with the image with artifacts to generate an image at "C2" that is free of or has at least less artifacts resulting from the high pitch operation, such that the image appears to have been acquired with regular pitch as shown at "A1" under full angular coverage for the slice indicated for example at "B2".

[0106] Then, Figure 6Two sets of corrected image slices are shown, corrected by the neural network trained as discussed above. The input image is shown at “II”, the corrected image is shown at “CI”, and the ground truth image is provided at “GTI”. Therefore, Figure 6 Examples of two image slices that have been corrected using the novel method described herein are shown. Aside from some minor degradation, the trained neural network is able to correct most artifacts in the images. It should be noted that the projection-domain-based network can be combined with the embodiments described herein, which effectively extrapolates the detector rows, thereby substantially improving Z-coverage.

[0107] Then, Figure 7 A detailed example of an imaging system that also corrects the image data discussed above is shown. System 100 includes a general fixed frame 102 and a rotating frame 104. System 100 is as follows: Figure 2 A specific embodiment of system 50 is shown. A rotating gantry 104 is rotatably supported by a fixed gantry 102 via bearings (not visible) and rotates about the examination area about the z-axis (i.e., the axis of rotation). A radiation source 108 (e.g., an X-ray tube) is supported by and rotates with the rotating gantry 104, emitting X-ray radiation. The radiation source 108 can be about... Figure 2 A specific example of the described radiation source 60. A radiation-sensitive detector array 110 is positioned opposite the radiation source 108 across an inspection region 106 at an angular arc, and detects radiation traversing the inspection region 106 and generates a signal (projection data) indicating that radiation. The illustrated radiation-sensitive detector array 110 comprises a two-dimensional (2D) array with multiple rows arranged about each other along the z-axis. The radiation detector array 110 can be as follows: Figure 2 The detector 70 shown is a specific example. The reconstructor 112 reconstructs the signal and generates volumetric image data indicating the inspected area 106. The reconstructor 112 is capable of being about... Figures 1-2 The described apparatus is a specific example of the processing unit 20. The reconstructor 112 can be implemented via hardware and / or software. For example, the reconstructor 112 can be implemented via a processor (e.g., a central processing unit or CPU, microprocessor, controller, etc.) configured to run computer-executable instructions, such as those stored, encoded, embedded, etc., on a computer-readable medium (e.g., a memory device excluding non-transient media), wherein running the instructions causes the processor to perform one or more of the actions described herein and / or another action.

[0108] A support 118 (e.g., a couch) supports the subject in the examination region 106 and can be used to position the subject with respect to the x-axis, y-axis, and / or z-axis before, during, and / or after scanning. A computing system serves as an operator console 120 and includes output devices (e.g., a display configured to display reconstructed images) and input devices (e.g., a keyboard, mouse, etc.). Software resident on the console 120 allows the operator to control the operation of the system 100, e.g., identify reconstruction algorithms, etc. The operator console 120 can be located in the same or a different room than the system 100. Figure 1 Particular examples of output units are described.

[0109] Then, Figure 8 Details are provided regarding the helical data acquisition path undertaken by the system 100. Figure 8 A helical path 304 (with a pitch "d") of the subject 302, an x-ray focal point or focal spot 306 of the radiation source 108, and the detector array 110 are depicted. For a particular slice location 300 having the same z-coordinate as the focal spot 306 and the center of the detector array 110 at that point in time, as the gantry 104 moves in the z-direction relative to the subject 302, the upper half 308 (or first plurality of rows) of the detector array 110 collects data that is temporally earlier, while the lower half 310 (or second plurality of rows) of the detector array 110 collects data that is temporally later for the same slice. The relative pitch then relates the pitch d to the detector height H det is associated. Thus, data can be acquired at a low pitch and used to train the neural network. The same data can then be manipulated in order to reduce the effective detector size H det , thereby creating data that was actually acquired at a "high pitch" and which can also be used to train the neural network.

[0110] In another exemplary embodiment, a computer program or a computer program element is provided that is characterized by being configured to execute the method steps of the method according to one of the preceding embodiments, on an appropriate apparatus or system.

[0111] A computer program element might therefore be stored in a computer unit, which can also be part of an embodiment. This computer program element can be configured to perform or induce a performing of the steps of the method described above. Moreover, the computer program element might be configured to operate the components of a corresponding apparatus and / or system. The computer program element can be configured to operate automatically and / or to execute the orders of a user. A computer program might be loaded into a working memory of a data processor. The data processor can thus be equipped to carry out the method according to one of the preceding embodiments.

[0112] This exemplary embodiment of the present application covers both computer programs which use the present application from the very beginning, and computer programs which convert existing programs to programs using the present application by means of an update.

[0113] Further on, the computer program element might be able to provide all the steps of a method for operating and / or monitoring the system as described above.

[0114] According to a further exemplary embodiment of the present application, a computer readable medium, such as a CD-ROM, a USB stick or the like, is presented wherein the computer readable medium has stored thereon the computer program element. This computer program element is presented by the preceding section.

[0115] A computer program can be stored and / or distributed on a suitable medium, such as an optical storage medium or a solid state storage medium supplied together with or as part of other hardware, but also by means of an electronic signal. Computer programs can be in many forms such as, for example, a compiled form, a binary form, a hexa- decimal form, an assembly form or a high level code form. The utilizable type of computer program will depend on the technical characteristics of the computer which is used.

[0116] The computer program element might therefore be present on a medium, e.g., a record medium, a computer readable medium, a sim card or a memory device, with further programming, e.g., an operating system or a related software.

[0117] It has to be noted that embodiments of the application effect to different aspects of the application. In particular, some embodiments are described with respect to method claims whereas other embodiments are described with respect to device claims. However, unless explicitly stated otherwise, any combination of features belonging to one type of subject matter with features belonging to another type of subject matter is considered to be disclosed within this application. However, all features can be combined to provide synergistic effects of more than the simple addition of the features.

[0118] While the application has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary only. The application is not limited to the disclosed embodiments. Other variations that are within the spirit of the present application will become readily apparent to those skilled in the art from reading the foregoing description. The scope of the application is accordingly indicated by the appended claims, rather than the foregoing description, and all changes that come within the meaning and range of equivalents are intended to be embraced therein.

[0119] In the claims, the term "comprising" does not exclude other elements or steps, and the wording "a" or "an" does not exclude a plurality. A single processor or other unit can fulfil 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 construed as limiting the scope.

Claims

1. An apparatus for correcting computed tomography (CT) X-ray data acquired at high relative pitch, the apparatus comprising: Input unit; Processing unit; as well as Output unit, The input unit is configured to provide the processing unit with CT X-ray data of a person's body parts acquired at high relative pitch. The processing unit is configured to use machine learning algorithms to determine CT slice reconstruction data of the person's body parts that have no high relative pitch operation reconstruction artifacts or have reduced high relative pitch operation reconstruction artifacts. The machine learning algorithm is trained based on CT slice reconstruction data, which includes first CT slice reconstruction data with high relative pitch reconstruction artifacts and second CT slice reconstruction data with no or fewer high relative pitch reconstruction artifacts. The CT X-ray data for the first CT slice reconstruction data is derived from the CT X-ray data for the second CT slice reconstruction data, processed by simulating a high relative pitch acquisition process. The output unit is configured to output the CT slice reconstruction data of the body part of the person.

2. The apparatus according to claim 1, wherein, The CT X-ray data for the second CT slice reconstruction data were acquired at a low relative pitch.

3. The apparatus according to any one of claims 1 or 2, wherein, The CT X-ray data for the reconstruction of the second CT slice includes full angular coverage data.

4. The apparatus according to any one of claims 1-2, wherein, The CTX data for the second CT slice reconstruction data includes CT X-ray data of one or more body parts of the test subject.

5. The apparatus according to any one of claims 1-2, wherein, The CT X-ray data for the second CT slice reconstruction data is processed by simulating a high relative pitch acquisition process in order to provide CT X-ray data for the first CT slice reconstruction data that is effectively acquired at a higher relative pitch than the relative pitch used to acquire the CT X-ray data for the second CT slice reconstruction data.

6. The apparatus according to any one of claims 1-2, wherein, The CT X-ray data for the second CT slice reconstruction data is processed by simulating a high relative pitch acquisition process to provide the following CT X-ray data for the first CT slice reconstruction data: the CT X-ray data for the first CT slice reconstruction data is simulated as being acquired by a resized detector, the resized detector being resized for the detector used to acquire the CT X-ray data for the second CT slice reconstruction data.

7. The apparatus according to claim 6, wherein, The simulation involves reducing the size of the detector in a direction parallel to the rotation axis of the X-ray imaging system used to acquire the CT X-ray data for reconstructing data from the second CT slice.

8. The apparatus according to any one of claims 1-2, wherein, The CTX data for the reconstruction of the first CT slice includes CT X-ray data from one or more test subjects.

9. The apparatus of claim 8, wherein training the machine learning algorithm comprises registering at least some of the second CT slice reconstruction data with the first CT slice reconstruction data reconstructed based on the CT X-ray data of one or more test subjects.

10. The apparatus according to claim 8, wherein, The CT X-ray data for the reconstruction of the first CT slice includes CT X-ray data of the body parts of the one or more test subjects.

11. An imaging system, comprising: X-ray source; X-ray detector; as well as Apparatus for correcting computed tomography (CT) X-ray data acquired at high relative pitch, according to any one of claims 1-9; The X-ray source and the X-ray detector are configured to rotate around the body parts of a person and acquire CT X-ray data at a high relative pitch; and The device is configured to output CT slice reconstruction data of the body parts of the person.

12. A method for correcting computed tomography (CT) X-ray data acquired at high relative pitch, the method comprising: a) Provide the processing unit with CT X-ray data of the body parts of the person acquired at high relative pitch; b) Using the processing unit, a machine learning algorithm is employed to determine CT slice reconstruction data of the person's body portion that is free of or has reduced high relative pitch reconstruction artifacts; wherein the machine learning algorithm is trained based on the CT slice reconstruction data, and wherein the CT slice reconstruction data includes first CT slice reconstruction data with high relative pitch reconstruction artifacts, and second CT slice reconstruction data with no high relative pitch reconstruction artifacts or with fewer or less severe high relative pitch reconstruction artifacts, wherein the method includes processing the CT X-ray data for the second CT slice reconstruction data by simulating a high relative pitch acquisition process to derive the CT X-ray data for the first CT slice reconstruction data; and c) The output unit outputs the CT slice reconstruction data of the body part of the person.

13. A computer program unit for controlling an apparatus according to any one of claims 1 to 2 and / or a system according to claim 11, the computer program unit being configured, when run by a processor, to perform the method according to claim 12.

Images