Method for scattered radiation correction

Abstract

The invention relates to a computer-implemented method for scattered radiation correction. The method includes: receiving a plurality of projection images depicting an object; and receiving or determining a virtual three-dimensional reference image depicting the object. The method includes determining at least one scattering model and subtracting the at least one scattering model from the projection image. A corrected projection image is determined. The method includes determining a corrected medical image from the corrected projection image, and comparing the corrected image to a virtual reference image. The method includes checking whether an abortion criterion is reached. If an abortion criterion is reached, the method includes providing a corrected medical image. If the abortion criterion is not reached, the method includes adjusting at least one scattering model and repeating the method steps based on the adjusted scattering model.

Description

Technical Field

[0001] The invention relates to a method for scattered radiation correction in three-dimensional medical images, a system configured to carry out the method, a computer program product and a computer-readable storage medium. Background Art

[0002] It is known to determine or reconstruct a three-dimensional medical image based on a plurality of projection images. In this case, the three-dimensional medical image can depict an object. For this purpose, the object is positioned between an X-ray source and a detector. The projection images depict the object from different directions. Such a three-dimensional medical image is therefore based on X-ray imaging of the object.

[0003] It is also known that scattered radiation has a great influence on the image quality of such three-dimensional medical images. Scattered radiation has a great influence on the image quality of three-dimensional medical images determined based on projection images detected with cone beam geometry. In cone beam geometry, the object is magnified on the detector in relation to its distance from the X-ray tube and its distance from the detector. Such projection images in X-ray imaging are typically detected with the aid of a C-arm. Without correction, scattered radiation may cause streak artifacts (English: streak artifact), cupping artifacts (Cupping Artefakten) and / or blurring artifacts (English: smearing artifact) in the three-dimensional medical image. The artifacts may make medical diagnosis based on three-dimensional medical images difficult or infeasible. For the reasons described, it is necessary to reduce and / or correct scattered radiation.

[0004] It is known to position an anti-scatter grating in front of the detector in order to filter out scattered radiation. However, direct radiation that is not scattered radiation is also filtered out by the anti-scatter grating. This may result in a higher dose being required to achieve the desired image quality.

[0005] Alternatively, it is known to provide the largest possible distance between the object and the detector in order to reduce scattered radiation in the projection image. However, due to the increase in the cone beam geometry, this results in a relatively small area in the object detected by the detector and requires a large and therefore expensive detector.

[0006] Alternatively, it is known to remove or reduce scattered radiation in projection images by applying a training function to the projection images. However, such training functions are prone to errors. This may cause medically important structures to be removed from the projection images and for these reasons important medical facts can no longer be identified or shown in the three-dimensional medical images.

[0007] Alternatively, it is known that the scattered radiation can be modeled by a solution of the Boltzmann transport equation and corrected in this way in the projection image. However, this method is also prone to errors.

[0008] Alternatively, it is known that the scattered radiation is described or modeled by a low-frequency function. Depending on the material depicted in the three-dimensional medical image and the actually measured or determined Heinz (abbreviation: HU) value, the function for modeling the scattered radiation can be adjusted iteratively. However, for this purpose, it is necessary to know the material of the object. In addition, the method is also prone to errors and is computationally expensive. Summary of the invention

[0009] It is therefore an object of the present invention to provide a method by means of which scattered radiation in three-dimensional medical images can be corrected reliably and without mechanical intervention in the measurement structure.

[0010] The object is achieved by a method according to the invention for performing scattered radiation correction in three-dimensional medical images, by an apparatus for performing scattered radiation correction in three-dimensional medical images, by a computer program product and by a computer-readable storage medium. Advantageous further developments are described in detail in the following description.

[0011] The solution according to the invention for the purpose is described below not only with respect to the claimed device, but also with respect to the claimed method. The features, advantages or alternative embodiments mentioned here can also be transferred to other claimed subjects, and vice versa. In other words, substantial embodiments (such as for devices) can also be improved by means of features described or claimed in conjunction with the method. The corresponding functional features of the method are constituted by corresponding substantial modules.

[0012] Regardless of the grammatical part of speech of a particular term, people with either male or female gender identity are included.

[0013] The invention relates to a method for performing scattered radiation correction in a three-dimensional medical image. The method comprises the method step of receiving a plurality of projection images. The projection images here depict an object. The method comprises the method step of receiving or determining a virtual three-dimensional reference image. Here, the virtual three-dimensional reference image depicts the object. The method comprises the method step of determining at least one scatter model. The method comprises the method step of subtracting at least one scatter model from a plurality of projection images. Here, a plurality of corrected projection images are determined. The method comprises the method step of determining a corrected three-dimensional medical image based on a plurality of corrected projection images. The method comprises the method step of comparing the corrected three-dimensional medical image with the virtual three-dimensional reference image, wherein a quality parameter is determined. The method comprises the method step of checking whether a termination criterion is reached based on the quality parameter. If the termination criterion is reached, the method comprises the method step of providing a corrected three-dimensional medical image. If the termination criterion is not reached, the method comprises the method step of adjusting the scatter model based on the quality parameter and repeating the method step based on the adjusted scatter model.

[0014] In the method step of receiving a plurality of projection images, the plurality of projection images are received by means of an interface. Here, the plurality of projection images depict an object from different directions. In other words, two projection images in each of the plurality of projection images depict an object from two different directions. Here, the object may be a person or an animal or an object. Here, the object may be, in particular, a part of a person or a part of an animal or a part of an object.

[0015] In this case, the projection image is detected, in particular, by means of X-ray imaging. In this case, the object is positioned between an X-ray tube or an X-ray source or a radiation source and an X-ray detector or a detector. This arrangement is referred to below as a measuring device. The X-ray radiation or the radiation emitted by the X-ray tube passes through the object, where it is attenuated in intensity depending on the material of the object and the energy of the radiation and is detected or detected by the detector.

[0016] Here, the X-ray tube and / or the detector are designed to be rotatable around the object within a certain angular range, which may be, for example, 90° or 100° or 120° or 150° or 180° or 210° or 240° or 270° or 300° or 330° or 360°.

[0017] The measuring device forms in particular a cone beam geometry. In other words, the object is magnified on the detector as a function of its distance from the X-ray tube and its distance from the detector.

[0018] In the method step of receiving or determining a virtual three-dimensional reference image, a virtual three-dimensional reference image is received by means of an interface or determined by means of a computing unit. In this case, the virtual three-dimensional reference image depicts the object three-dimensionally. In this case, the virtual three-dimensional reference image advantageously has no artifacts caused by scattered radiation, or artifacts caused by scattered radiation are reduced in the virtual three-dimensional reference image.

[0019] Based on the plurality of projection images, a three-dimensional medical image can be determined, which depicts the object three-dimensionally. The virtual three-dimensional reference image depicts the object similarly to the three-dimensional medical image. In other words, the three-dimensional medical image and the virtual three-dimensional reference image are registered.

[0020] In the method step of determining at least one scatter model, at least one scatter model is determined by means of a calculation unit. Alternatively, an initial scatter model can be called up from a database.

[0021] Each of the plurality of projection images comprises a plurality of pixels arranged in a matrix. Here, each pixel comprises a pixel value, which is related to the intensity of the X-ray radiation measured at the location of the pixel. At least one scatter model comprises a similar number of pixels arranged in a matrix of similar dimensions. Each pixel of the scatter model comprises a pixel value.

[0022] In the method step of subtracting at least one scatter model from a plurality of projection images, the scatter model is subtracted from each projection image by means of a calculation unit. Here, from each pixel value of a pixel of the projection image, a pixel value of a pixel of the scatter model is subtracted, which pixel value corresponds to the position of the pixel in the matrix of the projection image according to its position in the matrix.

[0023] In this process, a plurality of corrected projection images are determined.

[0024] In the method step of determining a corrected three-dimensional medical image, a corrected three-dimensional medical image is determined based on a plurality of corrected projection images. In this case, the corrected three-dimensional medical image is determined similarly to the three-dimensional medical image. In particular, the corrected three-dimensional medical image can be determined based on a plurality of corrected projection images by means of filtered back projection.

[0025] In the method step of comparing the corrected 3D medical image with the virtual 3D reference image, a quality parameter is determined by means of a computing unit. The quality parameter describes the similarity between the corrected 3D medical image and the virtual 3D reference image.

[0026] In the method step of checking whether the termination criterion is reached, it is determined whether the termination criterion is reached based on the quality parameter by means of a computing unit. The termination criterion can, for example, preset a target value for the quality parameter. The termination criterion is reached when the quality parameter is below the target value. Alternatively, the termination criterion can be reached when the quality parameter exceeds the target value. Alternatively or additionally, in an iterative method as described below, the termination criterion can preset a maximum number of iterations. Alternatively or additionally, the termination criterion can preset a minimum difference between two quality parameters from iterations that follow each other. If the difference between two quality parameters from two iterations that follow each other is less than the minimum difference, the termination criterion is reached.

[0027] Here, the suspension criterion may include one of the described suspension criteria or an “or” combination of one of the described suspension criteria.

[0028] If the abort criterion is reached, the method comprises the method step of providing a corrected three-dimensional medical image by means of an interface.

[0029] If the termination criterion is not reached, in the method step of adjusting at least one scatter model, the scatter model is adjusted by means of a calculation unit as a function of the quality parameter. In this case, it is particularly advantageous to adjust at least one scatter model so that when the quality parameter is re-determined, the quality parameter approaches the termination criterion. Subsequently, the method steps are repeated based on the adjusted scatter model. In this case, in particular, the method steps of subtracting at least one scatter model from a plurality of projection images, determining a corrected three-dimensional medical image, comparing the corrected three-dimensional medical image with a virtual three-dimensional medical image, and checking are performed again based on the adjusted scatter model.

[0030] In particular, the steps are carried out iteratively until a termination criterion is reached.

[0031] The inventors have recognized that scattered radiation in a plurality of projection images can be corrected based on a virtual three-dimensional reference image and that scattered radiation or artifacts resulting therefrom can also be corrected in the three-dimensional medical image in the described manner. The inventors have recognized that it is therefore not necessary to adjust the measurement structure and that the measurement structure can be optimized in this manner with respect to dose and amplification. The inventors have also recognized that no knowledge of the object is required with the described method. The inventors have recognized that with the described method it can be ensured that during the correction process medically important structures are not incorrectly removed from the three-dimensional medical image.

[0032] According to one aspect of the invention, the quality parameter is based on a structural similarity value.

[0033] In particular, the corrected 3D medical image and the virtual 3D reference image include a plurality of voxels arranged in a 3D matrix and each including a voxel value. Here, the 3D matrix of voxels of the corrected 3D medical image and the virtual 3D reference image is identical in its dimensions.

[0034] The structural similarity value is known as a structural similarity index (SSIM) and is determined based on the mean value and standard deviation of the voxel values ​​of the corrected 3D medical image and the virtual 3D reference image.

[0035] The inventors have realised that the structural similarity value is a good and standardized measure for assessing the quality of a rectified three-dimensional medical image.

[0036] According to an alternative optional aspect of the invention, the quality parameter is based on an alternative spacing measure.

[0037] For example, the quality parameter can be based on the mean square difference between the voxel values ​​of two corresponding voxels of the corrected three-dimensional medical image and the virtual three-dimensional reference image. In this case, the corresponding voxels are arranged at the same position in the three-dimensional matrix. In particular, the quality parameter can be the square root of the sum of the variances of all voxels.

[0038] The inventors have recognized that different measures are suitable for assessing the quality of a corrected three-dimensional medical image.The inventors have recognized that a desired measure may be used to determine the quality parameter depending on the computing power and the type of quality assessment.

[0039] According to another aspect of the invention, a scatter model is determined for each projection image of the plurality of projection images.

[0040] In particular, an individualized scatter model is determined for each of the projection images. In this case, the individualized scatter models can be at least partially identical. In particular, at least two scatter models of two different projection images can be different.

[0041] The inventors have recognized that different scatter models can be flexibly determined for different projection images using the method. The inventors have recognized that different situations that may occur when the X-ray tube and the detector are rotated around the object can be taken into account. In particular, different situations can be taken into account by different scatter models when different projection images are detected from different directions.

[0042] According to another aspect of the invention, at least one scattering model is based on a low frequency signal or on a basis function.

[0043] In particular, the low-frequency signal or low-frequency function can be as described in Trapp, P, Maier, J, Susenburger, M, Sawall, S, Kachelrie β, M. Empirical scatter correction: CBCT scatter artifact reduction without prior information. Med Phys. 2022; 1–19. https: / / doi.org / 10.1002 / mp.15656 is constructed as described in .

[0044] The basis functions may in particular be so-called basis splines or B-splines.

[0045] The inventors have recognized that the scatter model can model the scattered radiation using a simple function. The inventors have recognized that for this purpose, a standard function can initially be assumed. In this way, the same output or start can be selected for the scatter model for each measurement structure.

[0046] According to another aspect of the invention, the plurality of projection images are projection images detected during an intervention.

[0047] In other words, the projection images are detected during the intervention. In other words, the projection images are detected intra-interventionally.

[0048] Here, an intervention has been carried out or performed on the object. Here, the intervention can be, in particular, a medical intervention. The intervention can be, for example, a catheter examination or a stent operation or the like.

[0049] The inventors have recognized that the projection images detected during interventions in particular involve scattered radiation, since during interventions the dose should be as low as possible, the projection images should be detected as quickly as possible and the geometry of the measured structure is restricted. The inventors have recognized that this situation in particular requires subsequent scattered beam correction.

[0050] According to another aspect of the invention, the plurality of projection images depict the object from different directions in a cone beam geometry.

[0051] In other words, in order to record each projection image, the object is projected onto the detector from another direction. In other words, in each projection image, the object is oriented differently with respect to the detector and / or the X-ray tube.

[0052] Due to the cone beam geometry, the object is magnified on the detector as a function of its distance from the x-ray tube and its distance from the detector.

[0053] In particular, a plurality of projection images can be acquired with the aid of a C-arm.

[0054] The inventors have recognized that in particular projection images of cone-beam geometries involve scattered radiation. The inventors have also recognized that in order to determine a three-dimensional medical image, the object has to be projected from different directions.

[0055] According to another aspect of the invention, the virtual three-dimensional reference image is a computed tomography image.

[0056] A computed tomography image is acquired with the aid of a computed tomography system. The computed tomography image depicts the object.

[0057] In particular, a three-dimensional medical image of the object is determined based on a plurality of projection images. The computed tomography image is based on the three-dimensional medical image. In other words, the computed tomography image is superimposed on the three-dimensional medical image so that both depict the object in the same way or in the same orientation.

[0058] For this purpose, the method can in particular initially comprise the method step of registering the computed tomography image with the three-dimensional medical image, wherein a virtual three-dimensional reference image is determined.

[0059] Computed tomography images can be used pre-interventionally or be recorded or detected prior to intervention.

[0060] The inventors have recognized that a computed tomography image can be assumed to be virtually free of scattered radiation. The inventors have recognized that a computed tomography image can be used as a virtual three-dimensional reference image. The inventors have recognized that a computed tomography image of an object is usually detected pre-interventionally. The inventors have recognized that the computed tomography image can be used as a virtual three-dimensional reference image and that in this way no additional recording or creation of a reference image is required. The inventors have recognized that possibly different orientations of objects in the computed tomography image and the three-dimensional medical image can be compensated by registration.

[0061] According to a further aspect of the invention, a virtual three-dimensional reference image is determined based on the magnetic resonance tomography image.

[0062] The magnetic resonance tomography images are acquired here with the aid of a magnetic resonance tomography system. The magnetic resonance tomography images depict the object.

[0063] In particular, a three-dimensional medical image of the object is determined based on the plurality of projection images. In particular, the method can include for this purpose a method step of registering the magnetic resonance tomography image with the three-dimensional medical image, wherein a virtual three-dimensional reference image is determined.

[0064] The magnetic resonance tomography image comprises a plurality of voxels which are arranged in a matrix corresponding to the matrix of the (corrected) three-dimensional medical image or having the same dimensions. Each voxel of the matrix comprises a voxel value. When determining the virtual three-dimensional reference image, for each voxel a voxel value which is expected when a similar image of the object is detected by means of X-ray radiation is derived based on the voxel value of the magnetic resonance tomography image.

[0065] In particular, magnetic resonance tomography images can be acquired pre-interventionally.

[0066] The inventors have recognized that a virtual three-dimensional reference image can be determined based on magnetic resonance tomography images. The inventors have recognized that based on the voxel values ​​of the magnetic resonance tomography images, voxel values ​​to be expected in the image detected by means of X-ray radiation can be derived. The inventors have recognized that in this way, previously detected magnetic resonance tomography images can be used as a basis for a virtual three-dimensional reference image, so that the virtual three-dimensional reference image does not need to be additionally recorded. The inventors have recognized that by registration, the positioning of the object in the magnetic resonance tomography image that may deviate from the three-dimensional medical image can be corrected.

[0067] According to a further aspect of the invention, the method comprises the method step of applying a first training function to the magnetic resonance tomography image, wherein a virtual three-dimensional reference image is generated.

[0068] Typically, the training function mimics the cognitive functions that connect humans to the human mind. In particular, through training based on training data, the training function can adapt to new situations and recognize and extrapolate patterns.

[0069] Typically, the parameters of the training function can be adjusted by means of training. In particular, supervised training, semi-supervised training, unsupervised training, reinforcement learning and / or active learning can be used for this purpose. In addition, representation learning (an alternative term is "feature learning") can also be used. In particular, the parameters of the training function can be adjusted iteratively through a plurality of training steps.

[0070] In particular, the training function may include a neural network, a support vector machine, a random tree or a decision tree and / or a Bayesian network, and / or the training function may be based on k-means clustering, Q-learning, a genetic algorithm and / or an association rule. In particular, the training function may include a combination of multiple unrelated decision trees or a set of decision trees (random forest). In particular, the training function may be determined by means of XGBoosting (eXtreme Gradient Boosting). In particular, the neural network may be a deep neural network, a convolutional neural network or a convolutional deep neural network. In addition, the neural network may be an adversarial network, a deep adversarial network and / or a generative adversarial network. In particular, the neural network may be a recurrent neural network. In particular, the recurrent neural network may be a network with long short-term memory (LSTM), in particular a gated recurrent unit (GRU). In particular, the training function may include a combination of the methods described. In particular, the method for training a function described herein is referred to as a network architecture of the training function.

[0071] The first training function can be, in particular, a Transformer Network or a Vision Transformer Network or a Convolutional Network or a Multilayer Perceptron or a Generative Adversarial Network (GAN).

[0072] The first training function is trained in particular by providing magnetic resonance tomography images of a plurality of objects as training input data, to which the first training function is applied. In addition, corresponding three-dimensional images of the objects based on X-ray radiation are provided as training output data. Here, the images of the training input data correspond to the images of the training output data. The two images corresponding to each other are registered with each other and depict the same object in the same position. When the first training function is applied to the images of the training input data, an output image is determined. The output image is compared with the corresponding image of the training output data. The first training function is then adjusted so that when the first training function is reapplied, the image determined in this way better matches the corresponding image of the training output data.

[0073] The inventors have recognized that a virtual three-dimensional reference image can be determined by applying a first training function to a corresponding magnetic resonance tomography image. The inventors have recognized that an image determined by means of another imaging modality can be changed in this way so that the image is identical to a three-dimensional medical image detected by means of another imaging modality.

[0074] According to a further aspect of the invention, a virtual three-dimensional reference image is determined based on the magnetic resonance tomography image by means of a heuristic algorithm.

[0075] In particular, a heuristic algorithm is applied to the magnetic resonance tomography images and a virtual three-dimensional reference image is determined in the described manner. In particular, a heuristic algorithm known for the described purpose can be applied to the magnetic resonance tomography images for this purpose.

[0076] The inventors have recognized that, as an alternative to the first training function, a heuristic algorithm can be used in order to determine the virtual three-dimensional reference image.

[0077] According to another aspect of the present invention, a virtual three-dimensional reference image is determined based on a three-dimensional medical image. Here, the three-dimensional medical image is based on a plurality of projection images.

[0078] In this case, in particular, a virtual three-dimensional reference image is derived from the three-dimensional medical image. In this case, the three-dimensional medical image is corrected by a first estimated scattered radiation correction.

[0079] The inventors have realized that a virtual three-dimensional reference image can be determined based on the three-dimensional medical image. Therefore, there is no need to detect another medical image. In this way, time and costs can be saved. In addition, no additional radiation and thus a higher dose need to be applied to the object. The inventors have realized that if a virtual three-dimensional reference image is derived from the three-dimensional medical image, the registration can be abandoned. In this way, potential errors can also be avoided by reference.

[0080] According to another aspect of the invention, the method comprises the method step of receiving or determining a three-dimensional image. The method further comprises the method step of applying a second training function to the three-dimensional medical image, wherein a virtual three-dimensional reference image is determined.

[0081] In the method step of receiving a three-dimensional medical image, the three-dimensional medical image is received by means of an interface. Here, the three-dimensional medical image is based on a plurality of projection images. In particular, the three-dimensional medical image is determined or reconstructed based on the plurality of projection images as described above.

[0082] In an alternative method step of determining a three-dimensional medical image, the three-dimensional medical image is determined from a plurality of projection images by means of a computing unit. In particular, the three-dimensional medical image can be determined by means of filtered backprojection of the plurality of projection images.

[0083] In this case, a three-dimensional medical image depicts the object three-dimensionally.

[0084] The second training function can be designed in particular as described above with respect to the first training function. The second training function is in particular a generative adversarial network (abbreviation: GAN). The second training function is designed to remove scattered radiation in a three-dimensional medical image.

[0085] The second training function is trained in the following manner: three-dimensional medical image data of different objects are provided as training input data, and three-dimensional medical image data corrected for scattered radiation are provided as training output data. In this case, pairs depicting the same object are always formed in the images of the training input data and the images of the training output data. The second training function is applied to the training input data for training. The images determined in this way are respectively compared with the training output data belonging to the training input data in pairs. Based on the comparison, the second training function is adjusted so that when the second training function is reapplied, the images determined in this way better correspond to the training output data.

[0086] The inventors have recognized that for scattered radiation correction, known GANs can be used to create or determine a virtual three-dimensional reference image. In this case, it is particularly unimportant whether smaller medically relevant structures of the object are incorrectly represented in the virtual reference image determined in this way, since the virtual reference image does not correspond to the corrected three-dimensional medical image. The inventors have recognized that with the described method it is possible to ensure that all structures in the corrected three-dimensional medical image are correctly represented.

[0087] According to another aspect of the invention, when the quality parameter is below a target value, the abort criterion is reached.

[0088] Below the target value means in particular a particularly large similarity between the (corrected) three-dimensional medical image and the virtual reference image. The target value is predetermined. The quality parameter describes the difference between the virtual three-dimensional reference image and the corrected three-dimensional reference image.

[0089] Alternatively, when the quality parameter is defined in another way, the abort criterion may also be reached when a target value is exceeded.

[0090] Alternatively or additionally, the termination criterion may also be reached when a maximum number of iterations or passes of the method are performed as described above and / or when the quality parameter changes between iterations by less than a pre-determined threshold value.

[0091] In particular, different types of termination criteria can form a common termination criterion in that the method is terminated and a corrected three-dimensional medical image is provided when at least one of the conditions is met.

[0092] The inventors have recognized that the quality of the scattered radiation correction can be predetermined depending on a termination criterion. The inventors have recognized that it is possible to combine different termination criteria in order to avoid an infinite execution of the method if the quality parameter does not converge.

[0093] The invention also relates to a system for performing scattered radiation correction in three-dimensional medical images. The system comprises an interface and a computing unit. The interface and the computing unit are configured to perform the following method steps:

[0094] - receiving a plurality of projected images,

[0095] where the projected image depicts the object,

[0096] - receiving or determining a virtual three-dimensional reference image,

[0097] Wherein a virtual three-dimensional reference image depicts the object,

[0098] - determining at least one scattering model,

[0099] - subtracting at least one scattering model from the plurality of projection images,

[0100] wherein a plurality of corrected projection images are determined,

[0101] - determining a corrected three-dimensional medical image from the plurality of corrected projection images,

[0102] - comparing the corrected 3D medical image with a virtual 3D reference image,

[0103] The quality parameters are determined by

[0104] - Check whether the stop criteria are met based on quality parameters,

[0105] If the suspension criteria are met:

[0106] - then providing a corrected three-dimensional medical image,

[0107] If the suspension criteria are not met:

[0108] - then adjusting at least one scattering model according to the quality parameter,

[0109] - repeating the method steps based on the adapted scattering model.

[0110] Such a system may be designed in particular to carry out the method described above for scattered radiation correction in three-dimensional medical images and aspects thereof. The system is designed to carry out the method and aspects thereof in that the interface and the computing unit are designed to carry out the corresponding method steps.

[0111] The invention also relates to a computer program product with a computer program and to a computer-readable medium. A largely software-based implementation has the advantage that even systems that have been used to date can be retrofitted in a simple manner by means of a software update to operate in the described manner. In addition to the computer program, such a computer program product may include additional components such as documentation and / or additional components, as well as hardware components, such as hardware keys (software dongles, etc.), etc., in order to use the software.

[0112] In particular, the present invention also relates to a computer program product having a computer program, which can be directly loaded into a memory of a system, and which has program segments so that when the program segments are executed by the system, all steps and aspects of the method described above for performing scattered radiation correction in three-dimensional medical images are performed.

[0113] In particular, the present invention relates to a computer-readable storage medium on which program segments readable and executable by a system are stored so that when the program segments are executed by the system, all steps and aspects of the method described above for performing scattered radiation correction in three-dimensional medical images are performed.

[0114] The above characteristics, features and advantages of the present invention become clear and easy to understand in conjunction with the following drawings and descriptions. Here, the drawings and descriptions should not limit the present invention and its implementation methods in any way. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] In the different figures, identical parts are provided with corresponding reference numerals. The figures are generally not true to scale.

[0116] The accompanying drawings show:

[0117] Figure 1 A first exemplary embodiment of a method for scattered radiation correction in a three-dimensional medical image is shown.

[0118] Figure 2 A second exemplary embodiment of a method for scattered radiation correction in a three-dimensional medical image is shown.

[0119] Figure 3 A first exemplary embodiment of the method steps for determining a virtual three-dimensional reference image is shown.

[0120] Figure 4 A second exemplary embodiment of the method steps for determining a virtual three-dimensional reference image is shown.

[0121] Figure 5 A system for performing scattered radiation correction in three-dimensional medical images is shown. DETAILED DESCRIPTION

[0122] Figure 1 A first exemplary embodiment of a method for scattered radiation correction in a three-dimensional medical image MI is shown.

[0123] The method comprises the method step of receiving a plurality of projection images PI of REC-1. In this case, the plurality of projection images PI depicts an object. In this case, two projection images PI depict the object from two different directions, respectively. In particular, the projection image PI is a two-dimensional X-ray image of the object. In this case, the projection image PI can be detected, in particular, with the aid of a C-arm or an adjustable X-ray system or a computed tomography system. In particular, the plurality of projection images PI can be detected with a cone beam geometry. In this case, the object is arranged between a radiation source, in particular an X-ray source, and a detector, wherein the object is magnified on the detector in relation to its distance from the detector and the radiation source.

[0124] Here, the object can be a person or an animal or an object. In particular, the object can be a part of a person or an animal or an object.

[0125] In an optional method step of receiving REC-2 or determining DET-1, a three-dimensional medical image MI can be received or determined. Here, the three-dimensional medical image MI is based on a plurality of projection images PI. In particular, when determining DET-1, the three-dimensional medical image MI can be determined based on a plurality of projection images PI. For example, the three-dimensional medical image MI can be determined from a plurality of projection images PI by means of filtered back projection. Alternatively, as described, the three-dimensional medical image MI can be determined in advance based on a plurality of projection images PI. The three-dimensional medical image MI thus determined can then be received, for example, from a database. The three-dimensional medical image MI depicts the object three-dimensionally.

[0126] The method comprises a further method step of receiving REC-3 or determining DET-2 a virtual three-dimensional reference image VRI. The virtual three-dimensional reference image VRI depicts the object. In particular, the virtual three-dimensional reference image VRI and the three-dimensional medical image MI are registered with each other. In other words, the virtual three-dimensional reference image VRI and the three-dimensional medical image MI depict the object in the same orientation from the same viewing angle. Here, the virtual three-dimensional reference image VRI includes less scattered radiation than the three-dimensional medical image MI. Here, the virtual three-dimensional reference image VRI can be determined or detected in advance and then received for the method. Alternatively, a virtual three-dimensional reference image can be determined. For example, the virtual three-dimensional reference image VRI can be determined based on the three-dimensional medical image MI. Alternatively, the virtual three-dimensional reference image can be determined based on an alternative measurement of an image data set of the object. In particular, the virtual three-dimensional reference image VRI can be a computed tomography image of the object. In particular, the computed tomography image can be detected with the aid of a computed tomography system having as little noise as possible caused by scattered radiation. Alternatively, the virtual three-dimensional reference image VRI can be based on a magnetic resonance tomography image or a three-dimensional medical image MI.

[0127] The method comprises a further method step of determining DET-3 at least one scatter model SM. The scatter model SM describes an approximation of the scattered radiation. In particular, a dedicated scatter model SM can be determined for each projection image PI. In particular, the scatter model SM can be based on a low-frequency polynomial or signal or basis function.

[0128] The method comprises a further method step of subtracting SUB at least one scatter model SM from a plurality of projection images PI. If a separate scatter model SM is determined for each projection image PI, the corresponding scatter model SM is subtracted from each projection image PI. In this case, a plurality of corrected projection images are determined.

[0129] Each of the projection images PI comprises a plurality of pixels arranged in a matrix or pixel matrix. Here, each pixel comprises a pixel value. At least one scattering model SM also comprises a plurality of pixels arranged in a matrix, which corresponds in its dimensions to the matrix of the projection images PI. Each of the pixels also comprises a pixel value. When subtracting SUB, the pixel values ​​corresponding to their position in the matrix are subtracted from each other.

[0130] Furthermore, the method comprises the method step of determining DET-4 a corrected three-dimensional medical image MI from a plurality of corrected projection images. In particular, the corrected three-dimensional medical image MI is determined based on the corrected projection images, for example by means of filtered backprojection.

[0131] The method further comprises the method step of comparing the corrected three-dimensional medical image MI with a virtual three-dimensional reference image VRI COMP. In particular, both the corrected three-dimensional medical image MI and the virtual three-dimensional reference image VRI comprise a plurality of voxels arranged in a three-dimensional matrix or voxel matrix. The dimensions of the two voxel matrices are identical. In this case, each voxel in the voxel matrix comprises a voxel value. During the comparison, the voxel values ​​of two voxels that correspond to each other according to their position in the voxel matrix are compared with each other. A quality parameter is determined in this process.

[0132] The quality parameter can be related to the squared spacing of the voxel values, for example. Alternatively, the quality parameter can be based on a structural similarity value.

[0133] The method comprises a further method step of checking whether a termination criterion is reached based on a quality parameter. In particular, the termination criterion can be reached when the quality parameter exceeds or falls below a predefined value.

[0134] The method comprises the method step of providing a PROV corrected three-dimensional medical image MI when the termination criterion is reached. In this case, the corrected three-dimensional medical image MI can be provided in particular by means of an interface SYS.IF. In particular, the corrected three-dimensional medical image MI can be provided to a database and / or a display unit in order to display the three-dimensional medical image MI to a person.

[0135] If the abort criterion is not met, the method comprises the method steps of adjusting at least one scatter model SM of the ADA as a function of the quality parameter and repeating the method steps based on the adjusted scatter model SM.

[0136] In particular, when the method steps are repeated, the method steps of subtracting SUB, determining the DET-4 corrected three-dimensional medical image MI, comparing COMP and checking CHECK based on the adjusted scatter model SM are executed again. Based on the renewed checking, it is determined again whether the steps must be repeated with the aid of a re-adjusted scatter model SM or whether a termination criterion is reached and the corrected three-dimensional projection image determined in this way is provided.

[0137] Figure 2 A second exemplary embodiment of a method for scattered radiation correction in a three-dimensional medical image MI is shown.

[0138] Shown in Figure 1 The method steps of comparing COMP, checking CHECK, subtracting SUB and determining DET-4 corrected three-dimensional medical image MI are based on Figure 1 The description is composed of .

[0139] Figure 3A first exemplary embodiment of the method steps for determining a virtual three-dimensional reference image VRI of DET-2 is shown.

[0140] In this case, the virtual three-dimensional reference image VRI is based on a magnetic resonance tomography image. When determining the virtual three-dimensional reference image VRI of DET-2, a first training function is applied to the magnetic resonance tomography image in the method step of applying APP-1. In this case, a virtual three-dimensional reference image VRI is generated. In this case, the first training function is constructed as described above. In particular, the first training function can include a transformer network or a convolutional network (English: Convolutional Network) or a generative adversarial neural network or a multilayer perceptron. In particular, the first training function can be trained with the aid of reinforcement learning.

[0141] Alternatively, the virtual three-dimensional reference image VRI can be determined based on the magnetic resonance tomography images using a heuristic algorithm.

[0142] Figure 4 A second exemplary embodiment of the method steps for determining a virtual three-dimensional reference image VRI of DET-2 is shown.

[0143] In particular, the virtual three-dimensional reference image VRI is based on the three-dimensional medical image MI. In particular, the method steps of receiving REC-2 or determining DET-1 the three-dimensional medical image MI are not optional.

[0144] The method then comprises the method step of applying the second training function APP-2 to the three-dimensional medical image MI. A virtual three-dimensional reference image VRI is determined in this case. The second training function is designed as described above. In particular, the second training function comprises a generative adversarial neural network (English: GAN). In particular, the second training function is trained to remove noise or scattered radiation from the three-dimensional medical image MI.

[0145] Figure 5 A system SYS for scattered radiation correction in a three-dimensional medical image MI is shown.

[0146] The illustrated system SYS for scattered radiation correction in a three-dimensional medical image MI is designed to carry out the method according to the invention for scattered radiation correction in a three-dimensional medical image MI. The system SYS comprises an interface SYS.IF, a computing unit SYS.CU and a memory unit SYS.MU.

[0147] The system SYS can be in particular a computer, a microcontroller or an integrated circuit (IC). Alternatively, the system SYS can be a real or virtual computer network (the technical term for a real computer network is a "cluster" and the technical term for a virtual computer network is a "cloud"). The system SYS can be configured as a virtual system (the technical term is "virtualization") executed on a computer or a real computer network or a virtual computer network.

[0148] The interface SYS.IF can be a hardware interface or a software interface (e.g. a PCI bus, USB or Firewire). The computing unit SYS.CU can include hardware and / or software components, such as a microprocessor or a so-called FPGA (Field Programmable Gate Array). The memory unit SYS.MU can be designed as a non-permanent working working memory (Random Access Memory, RAM) or as a permanent mass storage device (hard disk, USB stick, SD card, solid state drive (SSD)).

[0149] The interface SYS.IF may in particular comprise a plurality of sub-interfaces, which carry out different method steps of the corresponding method according to the invention. In other words, the interface SYS.IF may be configured as a plurality of interfaces SYS.IF. The computing unit SYS.CU may in particular comprise a plurality of sub-computing units, which carry out different method steps of the corresponding method according to the invention. In other words, the computing unit SYS.CU may be configured as a plurality of computing units SYS.CU.

[0150] Where this is not explicitly stated, but is nevertheless meaningful and within the scope of the invention, individual embodiments, their individual sub-aspects or features can be combined or exchanged with one another without departing from the scope of the invention. Where transferable, the advantages of the invention described with reference to one embodiment also apply to other embodiments without explicit mention.

Claims

1. A computer-implemented method for performing scattered radiation correction in a three-dimensional medical image (MI), the method comprising the following method steps: - receiving (REC-1) a plurality of projection images (PI), wherein the projected image depicts an object, - receiving (REC-3) or determining (DET-2) a virtual three-dimensional reference image (VRI), wherein the virtual three-dimensional reference image (VRI) depicts the object, - determining (DET-3) at least one scattering model (SM), - subtracting (SUB) said at least one scatter model (SM) from said plurality of projection images (PI), wherein a plurality of corrected projection images are determined, - determining (DET-4) a corrected three-dimensional medical image (MI) from said plurality of corrected projection images, - comparing (COMP) said corrected three-dimensional medical image (MI) with said virtual three-dimensional reference image (VRI), The quality parameters are determined by - based on said quality parameters check whether the termination criteria are met, If the suspension criteria are met: - then providing (PROV) said corrected three-dimensional medical image (MI), If the suspension criteria are not met: - then adjusting (ADA) said at least one scattering model (SM) according to said quality parameter, - repeating the method steps based on the adapted scattering model (SM).

2. The method according to claim 1, Wherein the quality parameter is based on a structural similarity value.

3. The method according to any one of the preceding claims, Therein a scatter model (SM) is determined for each projection image (PI) of the plurality of projection images (PI).

4. The method according to any one of the preceding claims, Wherein the at least one scattering model (SM) is based on a low frequency signal or a basis function.

5. The method according to any one of the preceding claims, The plurality of projection images (PI) are projection images (PI) detected during the intervention.

6. The method according to any one of the preceding claims, The plurality of projection images (PI) depict the object from different directions in a cone-beam geometry.

7. The method according to any one of the preceding claims, The virtual three-dimensional reference image (VRI) is a computed tomography image.

8. The method according to any one of claims 1 to 6, In this case, the virtual three-dimensional reference image (VRI) is determined based on the magnetic resonance tomography image.

9. The method according to claim 8, The method further comprises the following method steps: - applying (APP-1) a first training function to the magnetic resonance tomography image, wherein the virtual three-dimensional reference image (VRI) is generated.

10. The method according to claim 8, The virtual three-dimensional reference image (VRI) is determined on the basis of the magnetic resonance tomography image by means of a heuristic algorithm.

11. The method according to any one of claims 1 to 6, wherein the virtual three-dimensional reference image (VRI) is determined based on the three-dimensional medical image, Wherein the three-dimensional medical image is based on the plurality of projection images.

12. The method according to claim 11, The method further comprises the following method steps: - receiving (REC-2) or determining (DET-1) said three-dimensional medical image, - applying (APP-2) a second training function to the three-dimensional medical image, wherein the virtual three-dimensional reference image (VRI) is determined.

13. The method according to any one of the preceding claims, When the quality parameter is lower than a target value, the termination criterion is reached.

14. A system (SYS) for performing scattered radiation correction in three-dimensional medical images, The system comprises an interface (SYS.IF) and a computing unit (SYS.CU), The interface (SYS.IF) and the computing unit (SYS.CU) are designed to carry out the following method steps: - receiving (REC-1) a plurality of projected images (PI), wherein the projected image depicts an object, - receiving (REC-3) or determining (DET-2) a virtual three-dimensional reference image (VRI), wherein the virtual three-dimensional reference image (VRI) depicts the object, - determining (DET-3) at least one scattering model (SM), - subtracting (SUB) said at least one scatter model (SM) from said plurality of projection images (PI), wherein a plurality of corrected projection images are determined, - determining (DET-4) a corrected three-dimensional medical image (MI) from said plurality of corrected projection images, - comparing (COMP) said corrected three-dimensional medical image (MI) with said virtual three-dimensional reference image (VRI), The quality parameters are determined by - based on said quality parameters check whether the termination criteria are met, If the suspension criteria are met: - then providing (PROV) said corrected three-dimensional medical image (MI), If the suspension criteria are not met: - then adjusting (ADA) said at least one scattering model (SM) according to said quality parameter, - repeating the method steps based on the adapted scattering model (SM).

15. A computer program product having a computer program, which can be directly loaded into a memory (SYS.MU) of a system (SYS), the computer program having program segments so that when the program segments are executed by the system (SYS), all steps of the method according to any one of claims 1 to 13 are performed.

16. A computer-readable storage medium on which a program segment readable and executable by a system (SYS) is stored, so that when the program segment is executed by the system (SYS), all steps of the method according to any one of claims 1 to 13 are performed.

Images