Providing classified data sets

Abstract

The invention relates to a method for providing a classification dataset, comprising: acquiring an image dataset of an examination object by means of a medical imaging device, the image dataset having a plurality of image points, each of which has a time intensity curve, the image points map an examination region of the examination object having at least one contrast-forming blood vessel segment, identifying first image points in the image data set as image points mapping the at least one contrast-forming blood vessel segment, the time intensity curve of the first image points having a predefined variability, providing a classification dataset on the basis of the image dataset and the first image point, the classification dataset having a classification between the first image point and another image point of the image dataset. The invention also relates to a medical imaging device and a computer program product.

Description

Technical Field

[0001] The present invention relates to a method for providing a classification dataset, a medical imaging device, and a computer program product. Background Technology

[0002] To obtain temporal changes in the examined area of ​​a subject, X-ray-based imaging methods are often used. The temporal and / or spatial changes to be acquired may include, for example, the diffusion and / or flow of contrast agents, especially contrast agent streams and / or contrast agent clumps, in hollow organs of the subject, such as vascular segments.

[0003] X-ray-based imaging methods typically involve digital subtraction angiography (DSA), in which at least two time-sequentially acquired X-ray images are subtracted from each other, these images at least partially mapping (or depicting) a common examination area. Furthermore, in DSA, a masking phase for acquiring at least one mask image and a filling phase for acquiring at least one filling image are often distinguished. The mask image can map the examination area without contrast, especially when no contrast agent is applied to the examination area. Conversely, the filling image can map the examination area with contrast, especially when a contrast agent is applied to the examination area. As a result of DSA, a differential image is typically provided by subtraction of the mask image and the filling image. Thus, components that are irrelevant to treatment and / or diagnosis and / or interfering, especially components that do not change over time, can usually be reduced and / or removed in the differential image.

[0004] Performing three-dimensional (3D) spatially resolved DSA (3D-DSA) is a common procedure for assessing cerebral blood vessels without interference from skeletal structures. Here, 3D-DSA is based on an initial mask run without contrast agent injection, specifically mask image acquisition. This is not optimal from a clinical or usability perspective. Movement of the subject between images can cause undesirable artifacts. Furthermore, contrast agent injection must be time-precise. Additionally, the acquisition time is doubled due to the masking and filling phases. Moreover, an additional X-ray dose is applied during the mask run. Summary of the Invention

[0005] Therefore, the technical problem to be solved by the present invention is to achieve efficient acquisition of contrast-forming (or contrast-enhanced) vascular segments in terms of time and X-ray dose.

[0006] This technical problem is solved according to the present invention through technical solutions of various individual aspects. Advantageous embodiments with suitable extended designs are given in the specification.

[0007] The invention relates, in a first aspect, to a method, particularly computer-implemented, for providing a classification dataset. In a first step, an image dataset of an examination object is acquired using a medical imaging device. The image dataset has multiple image points, each with a time-intensity curve. Furthermore, these image points are, in particular, time-resolvedly mapped to an examination region of the examination object having at least one contrasting vascular segment. In a further step, first image points in the image dataset are identified as image points mapping at least one contrasting vascular segment, the time-intensity curves of these first image points having predefined variability. In a further step, a classification dataset is provided based on the image dataset and the first image points. Here, the classification dataset has a classification between the first image points and other image points in the image dataset.

[0008] The subjects of examination may include, for example, human and / or animal patients and / or examination phantoms, especially vascular phantoms. The subjects of examination may have vascular segments, especially arteries or veins.

[0009] Acquiring the image dataset may, in particular, involve acquiring and / or reading from a computer-readable data storage device and / or receiving from a data storage unit, such as a database. Furthermore, the image dataset may be provided by a processing unit of a medical imaging apparatus for recording the image dataset. Advantageously, during the recording of the image dataset, a contrast agent is disposed in the at least one vascular segment. Advantageously, during the recording of the image dataset, the vascular segment has, in particular, complete and / or uniform, particularly homogeneous and / or consistent contrast agent filling. The medical imaging apparatus may, for example, include magnetic resonance imaging (MRT) and / or computed tomography (CT) and / or medical X-ray equipment, particularly medical C-arm X-ray equipment and / or ultrasound equipment and / or positron emission tomography (PET) equipment.

[0010] The image dataset may include a two-dimensional (2D) and / or three-dimensional (3D) spatially resolved mapping of the examination area of ​​the object, particularly at least one contrasting vascular segment. Furthermore, the image dataset may be time-resolved. Advantageously, the image dataset may map a contrast agent, particularly an X-ray-proof agent, disposed within at least one vascular segment as at least one contrasting vascular segment. The image dataset may have multiple image points, particularly pixels and / or voxels, each with a time-intensity curve. The time-intensity curve can map the temporal changes in image values, particularly intensity values ​​and / or attenuation values, of the corresponding image points. The examination area may include a spatial region of the object, such as a volume, particularly a region of interest containing at least one contrasting vascular segment.

[0011] Identifying first image points in an image dataset whose temporal intensity curves exhibit predefined variability, such as predefined frequencies and / or predefined frequency components, may include segmenting these first image points. Specifically, the identification of first image points may include variability analysis, particularly frequency analysis, of their respective temporal intensity curves. Here, at least one variability of the image values ​​of these image points, particularly at least one frequency and / or at least one frequency value, may be identified respectively. In particular, multiple variabilities of the image values ​​of these image points, particularly multiple frequencies and / or frequency values, may be identified respectively. Furthermore, the respective identified at least one variability, particularly the corresponding at least one frequency and / or at least one frequency value, may be compared with predefined variability, particularly predefined frequencies and / or predefined frequency values. In particular, the respective multiple variabilities, particularly multiple frequencies and / or frequency values, may be compared with predefined variability, particularly predefined frequencies and / or predefined frequency values. If the corresponding at least one variability, particularly at least one frequency and / or at least one frequency value, substantially matches the predefined variability, particularly predefined frequencies and / or predefined frequency values, then the corresponding image point can be identified as a first image point, particularly an image point mapping at least one contrasting vascular segment. In particular, if at least one of the corresponding plurality of variability, especially at least one of the corresponding plurality of frequencies and / or at least one of the corresponding plurality of frequency values ​​substantially coincides with a predefined variability, especially a predefined frequency and / or a predefined frequency value, then the corresponding image point can be identified as a first image point, particularly an image point mapping at least one contrasting vascular segment. The temporal intensity curve of the image point mapping at least one contrasting vascular segment in the image dataset can advantageously map the, in particular, non-rigid pulsation and / or periodic motion of at least one contrasting vascular segment. The temporal intensity curve of the image point mapping at least one contrasting vascular segment in the image dataset can also map additional, in particular physiological motions of the examined object, such as respiratory motion. Therefore, the temporal intensity curve of the image points in the image dataset can map a superposition of multiple, in particular physiological motions of the examined object. Furthermore, the temporal intensity curve of the image points in the image dataset can have a superposition of multiple frequencies corresponding to the respective mapped motions of the examined object. In contrast, the temporal intensity curve of the other image points, in particular image points not mapping at least one vascular segment, can map at least one motion of the examined object tissue that is different from pulsational motion, particularly physiological motion, such as respiratory motion.

[0012] Predefined variability, particularly predefined frequencies, can be given in advance or determined. According to the first variant, predefined variability can be given in advance, for example, based on user input. Alternatively or additionally, predefined variability can be determined based on physiological parameters, particularly the physiological parameters of the subject of examination. Physiological parameters may include, for example, the subject's heart rate, particularly pulse rate. Predefined variability can be characterized, for example, by one or more predefined frequencies and / or predefined amplitudes.

[0013] Providing a classification dataset may include storing it on a computer-readable storage medium and / or displaying it on a display unit and / or transmitting it to a processing unit. In particular, a graphical representation of the classification dataset can be displayed via the display unit.

[0014] Advantageously, a classification dataset can be provided based on the first image points and the additional image points in the image dataset. Here, the classification dataset may have classifications, such as annotations and / or masking, between the first image points and the additional, particularly the remaining, image points in the image dataset.

[0015] The proposed method advantageously enables the classification of image points, particularly at least the first and subsequent image points, from an image dataset without masking operations. The classification of image points can be advantageously based on the analysis of temporal variation patterns in the temporal intensity curves of the image points. This approach allows for the efficient acquisition of contrasting vascular segments in terms of both time and X-ray dose.

[0016] In another advantageous implementation of the proposed method, the predefined variability may include the heart rate of the subject being examined.

[0017] Advantageously, a predefined variability can be given in advance such that it includes, in particular, at least proportionally, the heart rate of the subjects being examined. Here, the predefined variability can be based on the statistical heart rates of multiple subjects or, in particular, the current or average heart rate of the subjects being examined.

[0018] The proposed implementation can advantageously improve the identification, and especially the classification, of image points in an image dataset that map at least one, particularly a segment of an artery and / or vein.

[0019] In another advantageous embodiment of the proposed method, a second image point with a substantially constant time-intensity curve can be identified as an image point in the image dataset that maps to at least one bone tissue. Here, a classification dataset may be additionally provided based on the second image point. Furthermore, the classification dataset may have classifications at least between the first and second image points.

[0020] The identification of the second image point can be implemented similarly to the identification of the first image point. The identification of the second image point may include segmenting the second image point. Specifically, the identification of the second image point may include variability analysis, particularly frequency analysis, of its respective time-intensity curve. At least one variability, particularly at least one frequency, and / or at least one frequency value of the corresponding image point's time-intensity curve can be compared with a variability threshold, particularly a frequency threshold. If the corresponding at least one variability, particularly at least one frequency, and / or the corresponding maximum frequency value is below the variability threshold, the corresponding image point can be identified as a second image point, particularly an image point mapping bone tissue.

[0021] Advantageously, the classification dataset may additionally be provided based on a second image point from the image dataset, and in particular, based on a first image point, a second image point, and the additional image points from the image dataset.

[0022] The classification dataset may have classification, such as annotation and / or masking, at least between the first and second image points in the image dataset. Advantageously, the classification dataset may also have classification between the first image point, the second image point, and additional, especially the remaining image points in the image dataset.

[0023] The proposed implementation can advantageously achieve efficient classification, and in particular differentiation, among multiple image points in an image dataset in terms of time and X-ray dose, which map at least one segment of blood vessel or bone tissue.

[0024] In another advantageous embodiment of the proposed method, providing the classification dataset may include providing a graphical representation based on the classification of image points. The graphical representation may include at least the first image points.

[0025] Advantageously, the provision of the graphic representation may include displaying the graphic representation via a display unit. The graphic representation may be provided according to the classification of image points, particularly first and additional image points. For example, the graphic representation may have color coding and / or annotation and / or masking according to the classification of image points. The graphic representation may advantageously include first image points, particularly image values ​​and / or time-intensity curves of the first image points. When providing the graphic representation, the additional image points may, for example, be masked and / or displayed at least partially transparently.

[0026] Especially for the 3D spatial resolution mapping of an area to be examined across multiple image points, the graphical representation can include virtual 2D projections of image points onto a virtual display plane.

[0027] The proposed implementation method can advantageously support medical operators in obtaining contrasting vascular segments in terms of both time and X-ray dose.

[0028] In another advantageous embodiment of the proposed method, the second image point can be excluded from the graphic representation.

[0029] Advantageously, the second image point can be excluded from the graphical representation, and in particular masked. Especially for the 3D spatial resolution mapping of the region being examined in the image point, the graphical representation can include a virtual 2D projection of the image point onto a virtual display plane. The second image point can be displayed at least partially, and in particular completely, transparently.

[0030] The proposed implementation can advantageously support medical operators in acquiring contrasting vascular segments, particularly those not obscured and / or overlapped by image points mapped from bone tissue, in terms of both time and X-ray dose.

[0031] In another advantageous embodiment of the proposed method, the medical imaging device can be designed as a medical X-ray device. Here, acquiring the image dataset may include acquiring projected images of the object being examined, particularly at least one contrasting vascular segment, via the X-ray device.

[0032] Advantageously, the X-ray device can have an X-ray source and an X-ray detector, particularly a flat panel detector and / or a linear array detector. The X-ray source and the X-ray detector can be arranged in a defined arrangement relative to each other, particularly opposite to each other. Furthermore, the defined arrangement of the X-ray source and the X-ray detector can be movable, particularly rotatable and / or translational, for example, supported relative to the object being examined. The X-ray source can be designed to emit X-rays to visualize the object being examined. In particular, the X-ray source can be designed to emit a cone-shaped or fan-shaped beam of X-rays to visualize the object being examined. The central ray and / or intermediate ray of the X-rays emitted by the X-ray source can define the projection direction. The X-ray detector can be designed to detect X-rays, particularly after interaction with the area of ​​the object being examined. Furthermore, the X-ray detector can be designed to provide a projected image based on the detected X-rays.

[0033] X-ray equipment can be designed, for example, as a C-arm X-ray device, an O-arm X-ray device, or a computed tomography (CT) device.

[0034] The proposed implementation method can advantageously achieve efficient acquisition of contrasting vascular segments in terms of time and X-ray dose using X-ray equipment.

[0035] In another advantageous embodiment of the proposed method, the motion field can be identified, and in particular reconstructed, based on the projected image. First image points can be additionally identified based on the motion field.

[0036] The motion field can have a particularly 2D or 3D spatial and temporal resolved representation, particularly a model, of the motion mapped in the projection image of at least one contrasting vascular segment, particularly the motion of the contrast agent disposed in at least one vascular segment. The motion field can be reconstructed based on the projection image. In particular, the reconstruction of the motion field can be achieved based on the projection image in a manner similar to 4D reconstruction. Furthermore, the motion field can be represented, particularly simulating, the positional changes of at least one contrasting vascular segment in a spatial and temporal resolved manner. The motion field can, for example, have a vector or tensor field representing the motion of at least one contrasting vascular segment. The motion represented, particularly simulated, of at least one contrasting vascular segment in the motion field can include translation and / or rotation and / or deformation of at least one contrasting vascular segment. Advantageously, a first image point can be additionally identified based on the motion field. The motion field can be advantageously determined based on the temporal intensity curves of image points in an image dataset.

[0037] In particular, the variability, especially the frequency, of the simulated motion can be identified at the image point level using the motion field. The identification of the first image point may include comparing the identified variability, especially the frequency, with predefined variability, especially the predefined frequency.

[0038] The proposed implementation can advantageously achieve reliable acquisition of contrasting vascular segments by means of motion mapped in the image dataset.

[0039] In another advantageous embodiment of the proposed method, the projected images can be mapped from at least partially different projection directions onto the object under examination, particularly at least one contrasting vascular segment. Here, the image dataset can be reconstructed from multiple projected images.

[0040] Advantageously, multiple projection images can be acquired using an X-ray device. These projection images map the object under inspection from at least partially, and particularly in completely different, projection directions and at least commonly map the inspection area. In particular, multiple projection images can be acquired around a common isocenter point. Advantageously, the defined arrangement of the X-ray source and X-ray detector can be positionable, and in particular movable, along a defined imaging trajectory in order to acquire, and in particular, to image, the multiple projection images.

[0041] Advantageously, an image dataset can be reconstructed from multiple projected images, for example, through particularly filtered backprojection reconstruction. In particular, the reconstruction can be performed similarly to 4D-DSA reconstruction. The image dataset can advantageously map the inspection region in 3D spatial and temporal resolution.

[0042] The proposed implementation can advantageously achieve efficient 3D acquisition of contrasting vascular segments in terms of time and X-ray dose.

[0043] In another advantageous embodiment of the proposed method, a restricted region can be identified in the projected image, which maps to a common inspection region of the objects being inspected. Here, the reconstruction of the image dataset can be limited to the restricted region.

[0044] The constrained region may include a spatially constraining volume and / or a spatially constraining surface, particularly at least partially within the object being examined. Advantageously, the constrained region may be identified as a spatial region of the object being examined, such as an overlapping region, mapped by each of a plurality of projection images. Identification of the constrained region may include segmentation of image points in the projection images, the segmentation being based, for example, on a particularly global threshold comparison, particularly a comparison of image values ​​of image points in the projection images with a pre-given threshold, and / or on a vesselness filter. Advantageously, the constrained region includes a mapping of at least one vascular segment of the object being examined, particularly additionally including a mapping of bone tissue of the object being examined.

[0045] Advantageously, the reconstruction of the image dataset can be limited to restricted regions, particularly the corresponding common spatial regions in the projected images.

[0046] The proposed implementation can advantageously achieve improved, particularly robust and / or artifact-free reconstruction of image datasets from projected images.

[0047] In another advantageous implementation of the proposed method, the recognition of the first image point can be based on machine learning.

[0048] Advantageously, the recognition of the first image point can be based on machine learning, particularly by applying a trained function to an image dataset as output data, especially the temporal intensity curves of image points in the image dataset. Here, the input data to the trained function can be based on the image dataset, particularly including the image dataset. Furthermore, a classification dataset, particularly for the recognition of the first image point, can be provided as the output data of the trained function.

[0049] The trained function can be trained using machine learning methods. In particular, the trained function can be a neural network, especially a convolutional neural network (CNN) or a network that includes convolutional layers.

[0050] A trained function maps input data to output data. The output data may also depend on one or more parameters of the trained function. These one or more parameters can be determined and / or adjusted through training. The determination and / or adjustment of one or more parameters of the trained function can be based, in particular, on data pairs consisting of training input data and associated training output data, especially comparative output data, wherein the trained function is applied to a provided classification dataset. Specifically, this determination and / or adjustment can be based on a comparison of the training mapped data with the training output data, especially comparative output data. A trainable function, i.e., a function with one or more parameters that have not yet been adjusted, is also commonly referred to as a trained function.

[0051] Other terms used for trained functions include trained mapping rules, mapping rules with trained parameters, functions with trained parameters, artificial intelligence-based algorithms, and machine learning algorithms. An example of a trained function is an artificial neural network, where the edge weights of the artificial neural network correspond to the parameters of the trained function. The term "neural network" can also be used instead of "neural network." In particular, a trained function can also be a deep artificial neural network. Another example of a trained function is a "support vector machine," and other machine learning algorithms can also be used as trained functions.

[0052] Advantageously, the trained function, especially the neural network, has an input layer and an output layer. In this case, the input layer of the trained function can be designed to receive input data. Furthermore, the output layer can be designed to provide mapping data, especially output data. Additionally, the input layer and / or output layer can each include multiple channels, especially neurons.

[0053] Advantageously, at least one parameter of the trained function can be adjusted based on a comparison of the classification training dataset with a comparative dataset, such as a segmented or annotated image dataset.

[0054] Furthermore, the trained function can be designed to identify second image points in an image dataset. The trained function can advantageously provide the identification of first and second image points as output data for classification datasets.

[0055] The proposed implementation can advantageously achieve improved, and particularly time-efficient, recognition of first image points in an image dataset, and especially the classification of image points in the image dataset.

[0056] In another advantageous embodiment of the proposed method, the identification of the first image point may additionally be based on the temporal intensity curves of image points in the adjacent region of the corresponding image point.

[0057] Adjacent regions may include one or more image points in an image dataset, which are directly adjacent to each other. Furthermore, adjacent regions may include multiple connected image points in an image dataset arranged within a predetermined spatial distance, particularly a radius, relative to each other, of the image points, wherein at least one of the multiple image points is directly adjacent to the corresponding image point.

[0058] Advantageously, motion, such as optical flow, can be determined using the temporal intensity curves of corresponding image points and image points in their adjacent regions. Determining optical flow based on the temporal intensity curves of corresponding image points and image points in their adjacent regions can include determining motion vectors, particularly vector fields. By adding image points in the adjacent regions of the corresponding image points, the mapped motion, particularly the direction of optical flow, and especially the flow direction, can also be identified based on the temporal intensity curves.

[0059] The proposed implementation can advantageously achieve improved recognition of the first image point.

[0060] In a second aspect, the present invention relates to a medical imaging apparatus designed to perform the proposed method for providing a classification dataset.

[0061] The advantages of the proposed imaging device substantially correspond to the advantages of the proposed method for providing classification datasets. The features, advantages, or alternative implementations mentioned herein can also be applied to other claimed technical solutions, and vice versa.

[0062] Medical imaging equipment may include, for example, magnetic resonance imaging (MRT) and / or computed tomography (CT) and / or medical X-ray equipment, particularly medical C-arm X-ray equipment and / or ultrasound equipment and / or positron emission tomography (PET).

[0063] Advantageously, the medical imaging device can be designed to acquire image datasets. Furthermore, the imaging device can have a processing unit and an interface. The processing unit can be designed to identify first image points in the image dataset. Additionally, the processing unit and / or the interface can be designed to provide a classification dataset.

[0064] In another advantageous embodiment of the proposed imaging device, the imaging device can be designed as a medical X-ray device. Furthermore, the X-ray device can be designed to acquire projected images of the object being examined.

[0065] Medical X-ray equipment can advantageously include a detector, particularly an X-ray detector, and a source, particularly an X-ray source, arranged in a defined configuration, such as on a C-arm. To acquire a projected image of the object being examined, the X-ray source can emit X-rays to visualize the object. The detector can be designed to acquire the X-rays after they interact with the object. The processing unit can acquire a projected image of the object using the detector based on the signal provided by the acquired X-rays.

[0066] In a third aspect, the present invention relates to a computer program product having a computer program that can be directly loaded into the memory of a processing unit, the computer program having program segments that, when the program segments are executed by the processing unit, perform all steps of a proposed method for providing a classification dataset.

[0067] The computer program product herein may include software having source code that still needs to be compiled and bound, or that only needs to be interpreted, or may include executable software code that only needs to be loaded into the processing unit for execution. This computer program product enables the rapid, repeatable, and robust execution, by means of the processing unit, of methods for providing classification datasets. The computer program product is configured to perform the method steps according to the invention via the processing unit.

[0068] Computer program products, for example, are stored on computer-readable storage media, or on a network or server, and can be loaded from the network or server into the processor of a processing unit, which can be directly connected to the processing unit or designed as part of the processing unit. Furthermore, control information of the computer program product can be stored on an electronically readable data carrier. The control information on the electronically readable data carrier can be designed such that when the data carrier is used in the processing unit, the control information executes the method according to the invention. Examples of electronically readable data carriers are DVDs, magnetic tapes, or USB flash drives, on which electronically readable control information, particularly software, is stored. When this control information is read from the data carrier and stored in the processing unit, all embodiments of the aforementioned method according to the invention can be implemented.

[0069] The largely software-based implementation has the following advantages: currently used processing units can be easily updated and modified via software to operate in accordance with the present invention. In addition to the computer program, such a computer program product may, if necessary, include additional components, such as documentation and / or additional parts, and hardware components, such as hardware keys (dongles, etc.) for software use. Attached Figure Description

[0070] Embodiments of the present invention are shown in the accompanying drawings and will be described in more detail below. In the different figures, the same features are represented by the same reference numerals. In the drawings:

[0071] Figures 1 to 4 The diagram illustrates different implementations of the proposed method for providing classification datasets.

[0072] Figure 5 A schematic diagram showing an exemplary time-intensity curve displaying image points of at least one vascular segment is provided.

[0073] Figure 6 A schematic diagram of a medical imaging device is shown. Detailed Implementation

[0074] Figure 1 A schematic diagram illustrating an advantageous implementation of the proposed method for providing a classification dataset is shown. In a first step, an image dataset of the subject to be examined using a medical imaging device (CAP-BD) is acquired. The image dataset may have multiple image points, each with a time-intensity curve. Furthermore, these image points may be mapped, in particular, at a time-resolved manner, to an examination region of the subject having at least one contrasting vascular segment. In a subsequent step, a first image point in the image dataset BD with a predefined variability in its time-intensity curve may be identified (ID-V) as an image point mapped to at least one contrasting vascular segment. In another step, a PROV-CDS classification dataset (CDS) may be provided based on the image dataset BD and the first image point. The classification dataset CDS may have classifications between the first image point and other image points in the image dataset BD. Advantageously, the predefined variability may include the heart rate of the subject. Advantageously, the identification (ID-V) of the first image point may be based on machine learning.

[0075] Furthermore, the identification ID-V of the first image point can be additionally based on the temporal intensity curve of image points in the adjacent region of the corresponding image point.

[0076] The provision of a classification dataset CDS, PROV-CDS, may include providing a graphical representation based on the classification of image points. The graphical representation may include at least the first image points. In particular, second image points may be excluded from the graphical representation, specifically masked.

[0077] Figure 2A schematic diagram illustrating another advantageous implementation of the proposed method for providing a classification dataset is shown. Here, a second image point with a substantially constant time-intensity curve can be identified as an image point in the image dataset BD that maps to at least one bone tissue, ID-B. Furthermore, a classification dataset CDS can be additionally provided based on the second image point, PROV-CDS. The classification dataset CDS can have at least one classification between the first and second image points.

[0078] Figure 3 A schematic diagram illustrating another advantageous implementation of the proposed method for providing a classification dataset is shown. The medical imaging apparatus can be designed as a medical X-ray device. Here, acquiring the image dataset CAP-BD can include acquiring a projected image PI of the object being examined via the X-ray device. Furthermore, a three-dimensional motion field ID-MF can be identified based on the projected image PI. Advantageously, a first image point can be additionally identified as ID-V based on the three-dimensional motion field MF.

[0079] Figure 4 A schematic diagram illustrating another advantageous implementation of the proposed method for providing a classification dataset is shown. Here, the projected images PI can map the objects to be inspected from at least partially different projection directions. Furthermore, the RECO-BD image dataset BD can be reconstructed from multiple projected images PI. Advantageously, restrictive regions can be identified in the projected images PI, which map to common inspection regions of the objects to be inspected. The reconstructed RECO-BD of the image dataset can be confined to these restrictive regions.

[0080] Figure 5 A schematic diagram of an exemplary time-intensity curve I(t) showing image points of at least one vascular segment is shown. Here, the time variation depicted by the time-intensity curve I(t), especially the intensity I, can be variable, especially in frequency, including predefined variability, particularly the heart rate of the subject being examined.

[0081] Figure 6A schematic diagram of a medical C-arm X-ray apparatus 37, including a processing unit PRVS, is shown as an example of a medical imaging device. The medical C-arm X-ray apparatus 37 may advantageously have a detector 34, particularly an X-ray detector, and a source 33, particularly an X-ray source, arranged in a defined configuration on a C-arm 38. The C-arm 38 of the C-arm X-ray apparatus 37 may be movably supported about one or more axes. To acquire a projected image PI of an examination subject 31 positioned on a patient support 32, the processing unit PRVS may send a signal 24 to the X-ray source 33. The X-ray source 33 may then emit an X-ray beam. When the X-ray beam is projected onto the surface of the detector 34 after interacting with the examination subject 31, the detector 34 may send a signal 21 to the processing unit PRVS. The processing unit PRVS may acquire the CAP-PI projected image PI using signal 21 and reconstruct the RECO-BD image dataset BD. The processing unit PRVS may also be designed to identify first image points in the ID-V image dataset BD.

[0082] The C-arm X-ray device 37 may also include an input unit 42, such as a keyboard, and a display unit 41, such as a monitor and / or display and / or projector. For example, in the case of a capacitive and / or resistive input display, the input unit 42 may preferably be integrated into the display unit 41. The input unit 42 may be advantageously designed to acquire user input. For this purpose, the input unit 42 may, for example, send a signal 26 to the processing unit PRVS. The processing unit PRVS may be designed to be controlled according to user input, particularly signal 26, and especially to execute methods for providing the PROV-CDS classification dataset CDS.

[0083] Display unit 41 can be advantageously designed to display a graphical representation of the classification dataset CDS. For this purpose, processing unit PRVS can send signal 25 to display unit 41.

[0084] The schematic diagrams included in the described figures do not depict any scale or size proportions.

[0085] Finally, it should be reiterated that the methods and apparatus described in detail above are merely embodiments, and those skilled in the art can modify these embodiments in different ways without departing from the scope of the invention. Furthermore, the use of the indefinite article "a" or "an" does not preclude the existence of multiple related features. Similarly, the terms "unit" and "element" do not preclude the existence of related components composed of multiple sub-components that function together, which may, if necessary, be spatially distributed.

[0086] In this application, the expression "based on" can be particularly understood as "using". In particular, the statement that the first feature is generated based on the second feature (alternatively: derived, determined, etc.) does not preclude the first feature from being generated based on the third feature (alternatively: derived, determined, etc.).

Claims

1. A method for providing a (PROV-CDS) classification dataset (CDS), comprising: - Acquire image datasets (BD) of the examined object (31) using medical imaging equipment (CAP-BD). The image dataset (BD) contains multiple image points, each with a time-intensity curve. These image points map the examination area of ​​the object (31) to which at least one vascular segment forms a contrast. - Identify the first image point in the ID-V image dataset (BD) as an image point mapping at least one contrasting vascular segment, where the temporal intensity curves of these first image points have predefined variability. -Based on the image dataset (BD) and the first image point provision (PROV-CDS) classification dataset (CDS), The classification dataset (CDS) has the classification between the first image point and the other image points in the image dataset (BD).

2. The method according to claim 1, characterized in that, The predefined variability includes the heart rate of the subject (31).

3. The method according to claim 1 or 2, characterized in that, The second image point with a substantially constant time-intensity curve is identified (ID-B) as an image point in the image dataset that maps to at least one bone tissue. The classification dataset (CDS) is also provided based on second image points (PROV-CDS). The classification dataset (CDS) has a classification between at least the first image point and the second image point in the image dataset (BD).

4. The method according to any one of the preceding claims, characterized in that, The provision of classification datasets (CDS) (PROV-CDS) includes providing graphical representations based on the classification of image points. The graphic representation includes at least the first image point.

5. The method according to claims 3 and 4, characterized in that, The second image point is excluded from the graphical representation.

6. The method according to any one of the preceding claims, characterized in that, The medical imaging equipment is designed as a medical X-ray device. The acquisition of the image dataset (BD) (CAP-BD) includes the acquisition of the projection image (PI) of the object (31) by means of an X-ray device (CAP-PI).

7. The method according to claim 6, characterized in that, Based on the projected image (PI) recognition (ID-MF) of the motion field, The first image point is also identified based on the motion field (ID-V).

8. The method according to claim 6 or 7, characterized in that, The projected image (PI) maps the inspected object from at least partially different projection directions (31). This includes an image dataset reconstructed from multiple projection images (PI) (RECO-BD).

9. The method according to claim 8, characterized in that, Restricted regions are identified in the projection image (PI), which map to the common inspection region of the objects being inspected (31). The reconstruction of the image dataset (BD) is restricted to the restricted area.

10. The method according to any one of the preceding claims, characterized in that, The identification of first image points (ID-V) is based on machine learning.

11. The method according to any one of the preceding claims, characterized in that, The identification of the first image point (ID-V) is also based on the temporal intensity curves of image points in the adjacent region of the corresponding image point.

12. A medical imaging device designed to perform the method according to any one of the preceding claims.

13. The medical imaging device according to claim 12, characterized in that, It is designed as a medical X-ray device, wherein the X-ray device is designed to acquire a projected image (PI) of the object being examined (31).

14. A computer program product having a computer program that can be directly loaded into a processing unit (PRVS) memory, the computer program having program segments that, when the program segments are executed by the processing unit (PRVS), perform all the steps of the method according to any one of claims 1 to 11.

Images