Provision of a radiation exposure distribution for an imaging system with a radiation source

Abstract

The invention relates to a computer-implemented method (M1) for the more accurate provision of data of a load distribution (16) of a radiation (6) from a radiation source (4) of an imaging system (1), comprising the steps: - Receiving setting data (10) of the imaging system (1); - Receiving a radiation exposure parameter (14) associated with the setting data (10); - Receiving data from the setting data (10) associated with an intensity distribution (15) of the radiation (6) at a radiation detector (3) of the imaging system (1); - Applying a trained first function (11) to the setting data (10), the load parameter (14) and the intensity distribution data (15) as first input data of the trained first function (11), whereby the load distribution data (16) are generated as first output data of the trained first function (11); - Providing the load distribution data (16); wherein the intensity distribution data (15) are provided by applying a trained second function (12) to the setting data (10) as the second input data of the trained second function (12), generating the intensity distribution data (15) as the second output data of the trained second function (12). The invention further relates to a data processing device (2), a computer-implemented training method (M2), and an imaging system (1).

Description

[0001] The invention relates to a computer-implemented method for providing data on the exposure distribution of radiation from a radiation source of an imaging system, as well as a computer-implemented training method for training a function, a corresponding data processing device, a computer program product and an imaging system.

[0002] For imaging systems with a radiation source that generates ionizing radiation, it is desirable, for example for regulatory reasons, to document the radiation exposure for patients and medical personnel as precisely as possible. A corresponding radiation exposure parameter can be measured during the operation of the imaging system for this purpose.

[0003] Methods for providing data on the exposure distribution of radiation from a radiation source are known in the prior art, e.g. from the publications DE 10 2019 215 242 A1, US 2023 / 0 080 631 A1, US 2023 / 0 181 142 A1 or US 2022 / 0 309 650 A1.

[0004] In X-ray-based imaging systems, such as X-ray angiography systems, a dose area product (DAP) can be measured as a radiation exposure parameter using a measuring device called a DAP chamber in the beam path of the imaging system. The measured DAP can be used not only to meet the aforementioned regulatory requirements but also to control the radiation source.

[0005] However, such a DAP (Dynamic Exposure Profile) can only indicate the cumulative exposure across the irradiated area, not its distribution across that area. In particular, it cannot be assumed that the exposure is evenly distributed across the area, i.e., homogeneously; rather, the radiation exposure may be inhomogeneously distributed. Specifically, the radiation source emits inhomogeneously distributed radiation onto the area, which can lead to an inhomogeneous distribution of this radiation across the irradiated area. Therefore, an estimate of the local radiation exposure, which can affect the skin, blood vessels, organs, or the like in X-ray-based angiography systems, is also incorrect based solely on the DAP. Rather, the local exposure may be higher than the hypothetically homogeneously distributed exposure based on the DAP.

[0006] The object of the invention is to determine more precisely the radiation load distribution from a radiation source of an imaging system.

[0007] The problem is solved by the subject matter of the independent patent claims. Advantageous embodiments of the invention are described by the dependent patent claims, the following description, and the figures.

[0008] The invention is based on the idea of ​​generating, at least by applying a trained function, data of the exposure distribution as first output data based on setting data of the imaging system, the exposure parameter and data of the intensity distribution of the radiation at a radiation detector.

[0009] According to one aspect of the invention, a computer-implemented method is provided for providing data of a load distribution of radiation from a radiation source, in particular a radiation source for generating ionizing radiation, to an imaging system.

[0010] The exposure distribution can also be referred to as the dose distribution. For example, exposure distribution data can be available as image data that depicts the distribution of exposure on a two-dimensional surface. A pixel value in the image can then indicate the local radiation exposure at that pixel.

[0011] In one step of the computer-implemented procedure, imaging system settings data is received. In a further step, a corresponding radiation exposure parameter or dose parameter is received. In a further step, corresponding radiation intensity distribution data from a radiation detector of the imaging system are received.

[0012] In a further step, a trained first function is applied to the received setting data, the received load parameter, and the received intensity distribution data as the first input data of the first function. The load distribution data is generated based on the input data and serves as the first output data of the trained first function. The generated load distribution data can then be made available in a subsequent step.

[0013] The provided data on the load distribution can be estimated, particularly by the trained function, without the need for a measurement to record the load distribution. Accordingly, such a measurement procedure can be advantageously dispensed with.

[0014] In particular, the load distribution data can be stored or at least temporarily stored on a suitable storage medium in the form of a corresponding computer-readable file, preferably a two-dimensional image file containing the two-dimensional intensity distribution. Providing the load distribution data can be accomplished, for example, by saving the computer-readable file to the storage medium.

[0015] For example, this data can be used for documentation purposes, such as in an electronic patient record. It is also possible that the load distribution data can be provided to an electronic system with a processing unit that can further process this data.

[0016] Unless otherwise specified, all steps of the computer-implemented method can be performed by a data processing device comprising at least one processing unit. In particular, the at least one processing unit is configured or adapted to perform the steps of the computer-implemented method. For this purpose, the at least one processing unit can, for example, store one or more computer programs, the execution of which causes the at least one processing unit to carry out the computer-implemented method.

[0017] This method allows for the provision of significantly improved data on radiation exposure distribution. In particular, the improved data prevents the underestimation of local radiation peaks that can affect specific areas, such as a patient's skin. For example, skin redness that may appear even weeks after radiation treatment at a localized area can be attributed to these radiation peaks based on the improved data. Furthermore, knowing the location of radiation peaks at specific areas of the irradiated surface allows for adjustments to the radiation treatment, thereby reducing side effects such as skin redness.

[0018] The imaging system settings, also known as the system configuration, can preferably be individually configured by a user of the imaging system, such as a physician or administrator. Specifically, these settings can be stored, or at least temporarily stored, on a suitable storage medium in the form of a corresponding computer-readable file containing the settings data. The settings can be retrieved, for example, by reading the computer-readable file from the storage medium.

[0019] The radiation exposure parameter is a quantity or value in dosimetry and the basis for calculating the radiation exposure of an irradiated body or area during exposure to a radiation source, for example, during an X-ray examination with an X-ray machine, such as fluoroscopy, angiography, or similar procedures. The radiation exposure parameter corresponds to the total radiation exposure reaching the irradiated area within the beam path of the radiation source. The unit of measurement for the radiation exposure parameter is cGy x cm². 2 or Gy × m 2 The radiation exposure parameter can, for example, correspond to a dose area product, DAP.

[0020] In X-ray-based imaging systems, such as X-ray-based angiography systems, the dose area product can be measured as a parameter of exposure using a measuring device called a DAP chamber in the beam path of the imaging system, which sums up the exposure over an irradiated area of ​​the radiation to a single value.

[0021] The exposure parameter is particularly dependent on the specific imaging system and its settings. If the settings change, the exposure parameter also changes. Therefore, the received exposure parameter corresponds to the one measured or determined by alternative means on the imaging system configured with the received settings. This exposure parameter can be stored, or at least temporarily stored, on a suitable storage medium in the form of a corresponding computer-readable file containing the exposure parameter. The exposure parameter can be received, for example, by reading the computer-readable file from the storage medium.

[0022] The received intensity distribution corresponds to that which was measured at the radiation detector in the imaging system set up with the received setting data, or which was determined at the radiation detector in an alternative way, i.e., not by measurement.

[0023] The intensity distribution data can correspond to a detector image of the radiation detector, which in particular graphically shows the intensity distribution of the radiation on an irradiated detector surface of the radiation detector. Specifically, this intensity distribution can be stored or at least temporarily stored on a suitable storage medium in the form of a corresponding computer-readable file, preferably a two-dimensional image file containing the two-dimensional intensity distribution. The intensity distribution can be retrieved, for example, by reading the computer-readable file from the storage medium.

[0024] In a less preferred embodiment, the intensity distribution can be provided by irradiating the radiation detector with the radiation from the radiation source under the settings of the imaging system, wherein there is no object or a homogeneous object in the beam path, and the radiation detector provides the intensity distribution as a two-dimensional detector image. A homogeneous object can be understood to be an object that does not influence the intensity distribution. For example, acrylic glass of a suitable thickness can be used for this purpose.

[0025] The settings data, the load parameter, and the intensity distribution data can now serve as input data for the trained first function. Preferably, the trained first function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. Specifically, the trained function can be based on a generic inverse algorithm. In particular, a computing unit can be configured to use the trained function as the first output data to generate the load data.

[0026] A KNN can be understood as a software code or a collection of several software code components, where the software code may include several software modules for different functions, for example one or more encoder modules and one or more decoder modules.

[0027] A KNN can be understood as a nonlinear model or algorithm that maps an input to an output, where the input is given by an input feature vector or an input sequence, and the output can be an output category for a classification task or a predicted sequence.

[0028] The KNN can, for example, be provided in a computer-readable form.

[0029] The neural network can comprise several modules, including a code module and a decoder module. These modules can be understood as software modules or corresponding parts of the neural network. A software module can be understood as software code that is functionally connected and combined with a unit. A software module can include or implement multiple processing steps and / or data structures.

[0030] The modules can, in particular, themselves represent neural networks or subnetworks. Unless otherwise specified, a module of the neural network can be understood to be a trainable and, in particular, a trained module of the neural network. For example, the neural network, and thus all its trainable modules, can be trained throughout before the method is carried out. In other implementations, however, different modules can be trained or pre-trained individually. In other words, the method according to the invention can correspond to a deployment phase of the trained, first function, in particular the KNN.

[0031] The trained first function can be trained in such a way that it is suitable for generating realistic and meaningful load distribution data as input data. Preferably, the trained first function can be trained and provided using a computer-implemented training method for training first functions.

[0032] By using the trained first function, the load distribution can be predicted with extreme precision and tailored to individual needs. Furthermore, using this trained first function, the processing unit can generate the load distribution data extremely quickly, robustly, and with lower computational power.

[0033] The trained, first function can be provided, in particular, as a general function for a series of identical or at least similar imaging systems, or applied to each specific imaging system of this series.

[0034] The proposed method can be used in particular during irradiation of a body using the imaging system, for example during a medical procedure, but also in material testing.

[0035] According to the invention, the intensity distribution data is provided by applying a trained second function to the setting data as the second input data of the trained second function, whereby the intensity distribution data is generated as the second output data of the second function.

[0036] This is advantageous at least insofar as no actual irradiation or measurement is required to determine the intensity distribution based on the settings data. Instead, the intensity distribution can be provided purely through computer implementation.

[0037] In particular, the second set of output data for the trained second function can thus form part of the first set of input data for the trained first function. The trained second function provides the intensity distribution as a digital, two-dimensional image, specifically as a corresponding image file.

[0038] The trained second function can be provided as a specific function for a single, specific imaging system within a series of identical or at least similar imaging systems, or applied only to the specific imaging system within that series. By linking the first trained function, which can generally be applied to the series of imaging systems, with the second trained function, the first trained function can essentially be specified for the particular imaging system.

[0039] In particular, the trained first and second functions are combined during application, for example, in a medical procedure with a patient in the beam. This offers a further advantage: when using this combination in the application, the processing unit only needs to receive the setting data and the load parameter to generate the load distribution. This significantly speeds up and simplifies the process.

[0040] The trained second function can be trained, in particular, to generate realistic and meaningful intensity distribution data as output data. Preferably, the trained second function can be trained and provided using a computer-implemented training method for training a second function. Preferably, the trained second function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. In particular, the trained second function can be based on a generic inverse algorithm. In particular, a computing unit can be configured to apply the trained second function to generate the intensity distribution as second output data.

[0041] According to at least one embodiment, the trained second function is provided by a computer-implemented second training procedure. The computer-implemented second training procedure includes, in particular, the following steps: - Receiving second input training data, including imaging system settings data; - Receiving second output training data, comprising intensity distribution data, wherein the second output training data are related to the second input training data; - Training a second function based on the second input training data and the second output training data by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data; - Deploying the trained, second function.

[0042] According to at least one embodiment, the load parameter is provided by applying a trained third function to the setting data as the third input data of the third function, whereby the load parameter is generated as the third output data of the third function.

[0043] This is advantageous at least insofar as no actual irradiation or actual measurement is required to determine the exposure parameter based on the setting data. Rather, the exposure parameter can be provided purely through computer implementation.

[0044] In particular, the third set of output data from the trained third function can thus form part of the first set of input data from the trained first function. The trained third function provides the load parameter as a value for this purpose.

[0045] The trained third function can, in particular, be provided as a specific function for a single, specific imaging system of a series of identical or at least similar imaging systems, or be applied only to the specific imaging system of this series.

[0046] A further advantage arises from combining the trained third function with the trained second function and the trained first function. When using this combination, the processing unit only needs to receive the configuration data so that it can generate the load distribution. This significantly speeds up and simplifies the process.

[0047] The trained third function can be trained, in particular, to generate realistic and meaningful load parameters as input data. Preferably, the trained third function can be trained and provided using a computer-implemented training method for training third functions. Preferably, the trained third function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. In particular, the trained third function can be based on a generic inverse algorithm. In particular, a computing unit can be configured to apply the trained third function as third input data to generate the load parameter.

[0048] According to at least one embodiment, the trained third function is provided by a computer-implemented third training procedure. The computer-implemented third training procedure includes, in particular, the following steps: - Receiving third-party input training data, including imaging system settings data, - Receiving third output training data, including load parameters, wherein the third output training data are related to the third input training data; - Training a third function based on the third input training data and the third output training data by generating third output data by applying the third function to the third input training data and adjusting parameters of the third function based on a comparison of the third output data with the third output training data; - Deploying the trained, third function.

[0049] According to at least one embodiment, the load characteristic is provided to be a dose area product or a quantity dependent on the dose area product.

[0050] For example, in this case, the radiation source is an X-ray tube. Analogous quantities also apply to other types of ionizing radiation.

[0051] According to at least one embodiment, the intensity distribution data is a two-dimensional X-ray projection image. For example, the intensity distribution can be represented by the color intensity of the individual pixels of the X-ray projection image.

[0052] The X-ray projection image, also known as the detector image, can graphically represent a two-dimensional array of pixels, where the pixel values ​​can represent measured values ​​from individual radiation sensors of the radiation detector.

[0053] According to at least one embodiment, the exposure distribution data is a two-dimensional image of the exposure distribution at an interventional reference point. This interventional reference point can correspond to the location where the body to be irradiated is to be placed in the beam path. Thus, the exposure distribution can precisely indicate the exposure that acts, or would act, on the body to be irradiated.

[0054] According to at least one embodiment, the setting data of the imaging system includes at least operating parameters of the radiation source and / or a collimator of the imaging system.

[0055] If the radiation source is, for example, an X-ray source, and in particular an X-ray tube, then the operating parameters of the radiation source may include, in particular, a peak kilovoltage (kVp), i.e., the maximum tube voltage applied to the X-ray tube when generating X-rays as ionizing radiation. The operating parameters of the radiation source may also include the tube current of the X-ray tube and / or the focal spot size, and so on.

[0056] A collimator is a physical barrier that focuses or aligns high-energy electromagnetic waves such as X-rays or gamma rays. For example, the spatial confinement of the beam from the radiation source can be specified or adjusted using the collimator's operating parameters.

[0057] According to at least one embodiment, the imaging system is an X-ray-based imaging system and the radiation source is an X-ray source, in particular an X-ray tube.

[0058] Another aspect of the invention relates to a computer-implemented first training method for providing a trained first function that generates data on the exposure distribution of radiation from a radiation source of an imaging system. This method can be executed, in particular, by a computing unit and comprises, in particular, the following steps.

[0059] In the first step of the procedure, initial input training data is received. This includes imaging system settings, radiation exposure parameters associated with these settings, and radiation intensity distribution data from a radiation detector of the imaging system.

[0060] In a second step of the process, initial baseline training data is received, which includes data on the load distribution. This baseline training data is related to the initial training data.

[0061] In a third step, a first function is trained based on the first input training data and the first output training data by generating first output data by applying the first function to the first input training data and adjusting parameters of the first function based on a comparison of the first output data with the first output training data.

[0062] In a fourth step, the trained, first function is made available.

[0063] The input training data can be received by a first training interface of an electronic training system, and the output training data by a second training interface of the electronic training system. The trained, first function can be provided by a third training interface of the electronic training system.

[0064] Preferably, the first function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. In particular, the function can be based on a generic inverse algorithm. In particular, the electronic training system can be configured to train the first function. The first function can be trained in such a way that, as a trained first function, it is suitable for generating or predicting realistic and meaningful data on the radiation exposure distribution as input data.

[0065] By comparing the initial data with the initial training data, the error of the function can be determined. For example, a so-called cost function can be calculated to quantify the error. The parameters of the function can then be adjusted using a known algorithm to minimize the cost function, particularly iteratively.

[0066] According to at least one embodiment of the computer-implemented first training method, the initial training data is provided by dosimetric measurements and / or dosimetric simulations.

[0067] The dosimetric measurements, serving as initial training data, must be performed using the corresponding input training data. A dosimetric measurement can be carried out, in particular, using a radiochromatic film. The film contains a dye that changes color when irradiated with ionizing radiation, allowing the degree of irradiation, the radiation profile, and the radiation exposure distribution to be characterized. The film can be digitized, particularly by scanning, and subsequently digitally processed to create a two-dimensional digital image.

[0068] Alternatively, the load distribution data can be provided as baseline training data through dosimetric simulation, based on the associated settings, load parameters, and intensity distribution. For example, a Monte Carlo simulation can be performed for this purpose.

[0069] The intensity distribution and / or the load parameter as input training data can also be provided based on measurements and / or simulations and / or by means of a trained function.

[0070] According to at least one embodiment of the computer-implemented first training method, it is provided that the trained first function is made available for a series of imaging systems.

[0071] The trained, first function can be provided, in particular, as a general function for a series of identical or at least similar imaging systems, or applied to each specific imaging system of this series.

[0072] According to at least one embodiment of the computer-implemented method for providing load distribution data, it is provided that the trained first function is provided by the computer-implemented first training method.

[0073] According to the invention, a computer-implemented second training method is provided for a trained second function that generates data on the intensity distribution of radiation at a radiation detector of an imaging system with a radiation source. This method can be executed, in particular, by a computing unit and comprises, in particular, the following steps.

[0074] In a first step of the procedure, second input training data can be received, which includes setting data of the imaging system.

[0075] In a second step of the process, a second set of output training data can be received, comprising data on the intensity distribution. This second set of output training data is related to the second set of input training data.

[0076] In a third step, a second function is trained based on the second input training data and the second output training data by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data.

[0077] In a fourth step, the trained, second function is made available.

[0078] The input training data can be received by a first training interface of an electronic training system, and the output training data by a second training interface of the electronic training system. The trained, second function can be provided by a third training interface of the electronic training system.

[0079] Preferably, the second function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. Specifically, the function can be based on a generic inverse algorithm. In particular, the electronic training system can be configured to train the second function. The second function can be trained in such a way that, as a trained second function, it is suitable for generating or predicting realistic and meaningful data on the intensity distribution of the radiation as input data.

[0080] By comparing the initial data with the initial training data, the error of the function can be determined. For example, a so-called cost function can be calculated to quantify the error. The parameters of the function can then be adjusted using a known algorithm to minimize the cost function, particularly iteratively.

[0081] According to at least one embodiment of the computer-implemented second training method, the trained second function is provided specifically for a particular imaging system within a series of imaging systems. The series of imaging systems may, in particular, comprise systems that are identical or at least structurally similar.

[0082] The second initial training data for training the specific second function can be provided, in particular, by means of the associated specific system. Specifically, the intensity distribution data of the specific system can be obtained as the second initial training data through measurements using the specific system, with no object or a homogeneous object in the beam path. Such measurements can be provided, or have been provided, in particular through regular, for example, annual service measurements.

[0083] In particular, it may be sufficient to train the second function with a relatively small amount of training data. This is because, in the computer-implemented procedure for providing load distribution data, the trained second function is linked to the trained first function. The trained second function serves only to specify the trained first function to the specific system of the associated trained second function, whereas the trained first function was trained with a relatively large amount of training data. Therefore, in this combination, the relatively small amount of training data is sufficient for training the second function.

[0084] According to at least one embodiment of the computer-implemented method for providing load distribution data, it is provided that the trained second function is provided by the computer-implemented second training method.

[0085] A further aspect of the invention relates to a computer-implemented, third training method for providing a trained, third function which generates data on a radiation exposure parameter from a radiation source of an imaging method. This method can be executed, in particular, by a computing unit and comprises, in particular, the following steps.

[0086] In a first step of the procedure, third input training data can be received, which includes setting data of the imaging system.

[0087] In a second step of the process, a second set of initial training data can be received, which includes load parameters. This third set of initial training data is related to the third set of initial training data.

[0088] In a third step, a third function is trained based on the third input training data and the third output training data by generating third output data through the application of the third function to the third input training data and adjusting parameters of the third function based on a comparison of the third output data with the third output training data.

[0089] In a fourth step, the trained, third function is made available.

[0090] The input training data can be received by a first training interface of an electronic training system, and the output training data by a second training interface of the electronic training system. The trained, third function can be provided by a third training interface of the electronic training system.

[0091] Preferably, the third function can be based on an artificial neural network (ANN), in particular with at least one convolutional plane. Specifically, the function can be based on a generic inverse algorithm. In particular, the electronic training system can be configured to train the second function. The second function can be trained in such a way that, as a trained second function, it is suitable for generating or predicting realistic and meaningful data on the intensity distribution of the radiation as input data.

[0092] By comparing the initial data with the initial training data, the error of the function can be determined. For example, a so-called cost function can be calculated to quantify the error. The parameters of the function can then be adjusted using a known algorithm to minimize the cost function, particularly iteratively.

[0093] According to at least one embodiment of the computer-implemented third training method, the trained third function is provided specifically for a particular imaging system within a series of imaging systems. The series of imaging systems may, in particular, comprise systems that are identical or at least structurally similar.

[0094] The third set of baseline training data for training the specific third function can be provided, in particular, by means of the associated specific system. Specifically, the load parameters of the specific system can be obtained as the third set of baseline training data through measurements using the specific system.

[0095] According to at least one embodiment of the computer-implemented method for providing load distribution data, it is provided that the trained third function is provided by the computer-implemented third training method.

[0096] Another aspect of the invention relates to a data processing device with at least one computing unit, which is adapted to carry out an inventive, computer-implemented method for providing data on the exposure distribution of radiation from a radiation source of an imaging system.

[0097] In particular, the data processing device for providing data on the exposure distribution of radiation from a radiation source of an imaging system is specified, comprising: - a first interface for receiving settings data from the imaging system; - a second interface for receiving a radiation exposure parameter associated with the setting data; - a third interface for receiving data relating to the setting data, namely an intensity distribution of the radiation at a radiation detector of the imaging system; - a computing unit for applying a trained, first function to the setting data, the load parameter and the intensity distribution data as first input data of the first function, wherein the load distribution data are generated as first output data of the first function; - a fourth interface for providing the load distribution data.

[0098] A computing unit can be understood, in particular, as a data processing device containing a processing circuit. The computing unit can therefore process data to perform arithmetic operations. This may also include operations to perform indexed access to a data structure, such as a lookup table (LUT).

[0099] The computing unit may, in particular, contain one or more computers, one or more microcontrollers, and / or one or more integrated circuits, for example, one or more application-specific integrated circuits (ASICs), one or more field-programmable gate arrays (FPGAs), and / or one or more systems on a chip (SoCs). The computing unit may also contain one or more processors, for example, one or more microprocessors, one or more central processing units (CPUs), one or more graphics processing units (GPUs), and / or one or more signal processors, in particular one or more digital signal processors (DSPs). The computing unit may also include a physical or virtual array of computers or other units of the aforementioned type.

[0100] In various embodiments, the computing unit includes one or more hardware and / or software interfaces and / or one or more storage units.

[0101] A storage unit can be volatile data storage, for example as dynamic random access memory (DRAM) or static random access memory (SRAM), or as non-volatile data storage, for example as read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory or flash EEPROM, ferroelectric random access memory (FRAM), or magnetoresistive random access memory.It can be designed as MRAM (magnetoresistive random access memory) or as phase-change random access memory, PCRAM (phase-change random access memory).

[0102] Further embodiments of the data processing device according to the invention follow directly from the various configurations of the associated method according to the invention, and vice versa. In particular, individual features and corresponding explanations as well as advantages relating to the various embodiments of the method according to the invention can be transferred analogously to corresponding embodiments of the data processing device according to the invention.

[0103] A further aspect of the invention provides a computer program with instructions. When the instructions are executed by at least one data processing device, in particular a data processing device according to the invention, the instructions cause the data processing device to carry out the inventive method for providing data on the exposure distribution of radiation from a radiation source of an imaging system.

[0104] The instructions can be provided, for example, as program code. This program code can be provided, for example, as binary code or assembly language, and / or as source code in a programming language such as C, and / or as a program script, such as Python.

[0105] Another aspect of the invention relates to an electronically readable data carrier comprising instructions which, when the data carrier is executed on at least one data processing device, in particular on a data processing device according to the invention, cause the data processing device to carry out the inventive method for providing data of a load distribution of radiation from a radiation source of an imaging system.

[0106] The computer program and the computer-readable storage medium are each computer program products containing the commands.

[0107] According to a further aspect of the invention, an imaging system is provided. The imaging system comprises a data processing device according to the invention, a radiation source for generating, in particular, ionizing radiation, and a radiation detector for detecting parts of the radiation, in particular parts of the ionizing radiation that have passed through an object to be imaged.

[0108] Another aspect of the invention relates to a first electronic training system for providing a trained, first function which generates data of a load distribution of radiation from a radiation source of an imaging system, comprising: - a first training interface for receiving initial input training data, comprehensive - Imaging system settings data, - radiation exposure parameters associated with the setting data, and - Data relating to the settings data of an intensity distribution of the radiation at a radiation detector of the imaging system depending on the settings data; - a second training interface for receiving initial output training data, including load distribution data, where the output training data is related to the input training data; - a computing unit for training a first function based on the first input training data and the first output training data, by generating first output data by applying the first function to the first input training data and adjusting parameters of the first function based on a comparison of the first output data with the first output training data; - a third training interface to provide the trained, first function.

[0109] The electronic first training system is specifically designed to carry out the computer-implemented first training procedure.

[0110] Further embodiments of the training system according to the invention follow directly from the various configurations of the training method according to the invention, and vice versa. In particular, individual features and corresponding explanations as well as advantages relating to the various configurations of the training method according to the invention can be transferred analogously to corresponding configurations of the training system according to the invention.

[0111] Another aspect of the invention relates to a second electronic training system for providing a trained, second function which generates data of an intensity distribution of radiation at a radiation detector of an imaging system with a radiation source, comprising: - a fourth training interface for receiving second input training data, including imaging system settings data; - a fifth training interface for receiving second output training data, comprising intensity distribution data, wherein the output training data are related to the input training data; - a computing unit for training a second function based on the second input training data and the second output training data, by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data; - a sixth training interface to provide the trained, second function.

[0112] The electronic, second training system is specifically designed to carry out the computer-implemented, second training procedure.

[0113] Further embodiments of the training system according to the invention follow directly from the various configurations of the training method according to the invention, and vice versa. In particular, individual features and corresponding explanations as well as advantages relating to the various configurations of the training method according to the invention can be transferred analogously to corresponding configurations of the training system according to the invention.

[0114] Another aspect of the invention relates to a third electronic training system for providing a trained, third function which generates a load characteristic of radiation from a radiation source of an imaging system, comprising: - a seventh training interface for receiving third input training data, including imaging system settings data, - an eighth training interface for receiving initial baseline training data, including load parameters, wherein the baseline training data are related to the input training data; - a computing unit for training a third function based on the third input training data and the third output training data, by generating third output data by applying the third function to the third input training data and adjusting parameters of the third function based on a comparison of the third output data with the third output training data; - a ninth training interface to provide the trained, third function.

[0115] The electronic third training system is specifically designed to carry out the computer-implemented third training procedure.

[0116] Further embodiments of the training system according to the invention follow directly from the various configurations of the training method according to the invention, and vice versa. In particular, individual features and corresponding explanations as well as advantages relating to the various configurations of the training method according to the invention can be transferred analogously to corresponding configurations of the training system according to the invention.

[0117] Further aspects of the invention relate to computer training programs comprising commands which, when the respective computer program is executed on an electronic training system, cause the electronic training system to carry out a respective training method according to the invention.

[0118] Further aspects of the invention relate to electronically readable training data carriers comprising commands which, when the respective data carrier is executed on an electronic training system, cause the electronic training system to carry out a respective training method according to the invention.

[0119] For use cases or application situations that may arise during the described procedures or training procedures and that are not explicitly described here, it may be necessary to issue an error message and / or a request for user feedback and / or to set a default setting and / or a predetermined initial state according to the procedure.

[0120] Further features and combinations of features of the invention will become apparent from the figures and their description, as well as from the claims. In particular, further embodiments of the invention need not necessarily include all features of any one of the claims. Further embodiments of the invention may have features or combinations of features not mentioned in the claims.

[0121] The invention is explained in more detail below with reference to specific embodiments and associated schematic drawings. In the figures, identical or functionally equivalent elements may be designated with the same reference numerals. The description of identical or functionally equivalent elements is not necessarily repeated with respect to different figures.

[0122] The figures show: Fig. 1 a schematic representation of an exemplary embodiment of an imaging system according to the invention; Fig. 2 a schematic flowchart of an exemplary embodiment of a computer-implemented method according to the invention for providing data of a load distribution; Fig. 3 a schematic representation of an exemplary embodiment of the computer-implemented method according to the invention for providing data of a load distribution; Fig. 4 a schematic flowchart of an exemplary embodiment of a computer-implemented first training method according to the invention for providing a trained first function; Fig. 5 a schematic flowchart of an exemplary embodiment of a computer-implemented second training method according to the invention for providing a trained second function; Fig. 6 a schematic flowchart of an exemplary embodiment of a computer-implemented third training method according to the invention for providing a trained third function; Fig. 7 a schematic diagram of an exemplary implementation of a trained, first function; Fig. 8 a schematic diagram of an exemplary implementation of a trained, second function; Fig. 9 a schematic representation of a first embodiment of an artificial neural network (ANN) of a trained function; Fig. 10 a schematic representation of a first embodiment of a foldable KNN of a trained function.

[0123] In Fig. Figure 1 schematically illustrates an exemplary embodiment of an imaging system 1 according to the invention. The imaging system 1 comprises at least one data processing device 2 and an imaging modality, which includes a radiation source 4 for generating ionizing radiation 6 and a radiation detector 3 for detecting portions of the ionizing radiation 6. For example, the imaging modality is configured as an X-ray-based imaging modality, such as an X-ray-based angiography system, and includes an X-ray source as the radiation source 4 and an X-ray detector as the radiation detector 3. Furthermore, a patient is shown as the object 5 to be imaged, which is located at an interventional reference point 9 of the imaging system 1.

[0124] The data processing device 2 has at least one computing unit 7, 17. The at least one computing unit 7 is configured to perform a computer-implemented method M1 according to the invention for providing data on the radiation exposure distribution of the radiation 6 from the radiation source 4 of the imaging system 1. A schematic flowchart of an exemplary embodiment of such a method M1 is shown in Fig. 2 is shown. For this purpose, a corresponding computer program product 8 can be executed on the data processing device 2.

[0125] In Fig. 2 is a schematic flowchart and in Fig. Figure 3 shows a schematic representation of an exemplary embodiment of a computer-implemented method M1 according to the invention for providing data on a load distribution 16, which can be carried out in particular by the computing unit 7, especially the data processing device 2. Fig. 2 and Fig. 3 will be discussed together below.

[0126] In a first step S1 of the procedure M1, setting data 10 of the imaging system 1 can be received by the processing unit 7. In one embodiment, this setting data 10 of the imaging system 1 can include at least operating parameters of the radiation source 4 and / or a collimator of the imaging system 1.

[0127] In a second step S2 of the procedure M1, data relating to the setting data 10 of an intensity distribution 15 of the radiation 6 at a radiation detector 3 of the imaging system 1 can be received by the processing unit 7. For example, the intensity distribution data 15 can represent a detector image, in particular a two-dimensional X-ray projection image.

[0128] Preferably, the intensity distribution data 15 can be provided by applying a trained second function 12 to the setting data 10 as the second input data of the trained second function 12, whereby the intensity distribution data 15 are generated as the second output data of the trained second function 12. The trained second function 12 can, in particular, be applied by the further processing unit 17, wherein the further processing unit 17 and the processing unit 7 can be configured as a joint processing unit 7.

[0129] In a third step S3 of the procedure M1, a radiation exposure parameter 14 associated with the setting data 10 can be received by the processing unit 7. For example, the radiation exposure parameter 14 can be a dose area product (DAP) or a quantity dependent on the dose area product.

[0130] Preferably, the load parameter 14 can be provided by applying a trained third function 13 to the setting data 10 as the third input data of the trained third function 13, whereby the load parameter 14 is generated as the third output data of the trained third function 13. The trained third function 13 can, in particular, be applied by the further processing unit 17, wherein the further processing unit 17 and the processing unit 7 can be configured as a joint processing unit 7.

[0131] Alternatively, the stress parameter can be measured using a DAP chamber in the beam path of the imaging system.

[0132] In a fourth step S4 of the procedure M1, a trained first function 11 can be applied to the setting data 10, the load parameter 14 and the intensity distribution data 15 as first input data of the trained first function 11, whereby the load distribution data 16 are generated as first output data of the trained first function.

[0133] The load distribution data 16 can in particular represent a two-dimensional image of the load distribution 16 at the interventional reference point 9 of the imaging system 1.

[0134] In a fifth step S5, the data of the load distribution 16 are provided.

[0135] Fig. Figure 4 shows a schematic flowchart of an exemplary embodiment of a computer-implemented first training method M2 according to the invention for providing a trained first function 11, which can be carried out, for example, by a first electronic training system 18 with at least one computing unit.

[0136] In a first step S21, initial input training data are received, which include setting data 10' of an imaging system 1, load parameters 14' of the radiation 6 associated with the setting data (10'), and intensity distribution data 15' of the radiation 6 associated with the setting data 10' at a radiation detector 3 of the imaging system 1.

[0137] In a second step S22, initial baseline training data are received, comprising data of the load distribution (16'), whereby the initial baseline training data are related to the initial input training data.

[0138] In a third step S23, a first function is trained based on the first input training data and the first output training data by generating first output data by applying the first function to the first input training data and adjusting parameters of the first function based on a comparison of the first output data with the first output training data.

[0139] In a fourth step S24, the trained, first function 11 is provided.

[0140] Fig. Figure 5 shows a schematic flowchart of an exemplary embodiment of a computer-implemented second training method M3 according to the invention for providing a trained second function 12, which can be carried out, for example, by a second electronic training system 19 with at least one computing unit.

[0141] In a first step, S31 receives second input training data, which includes setting data 10' of an imaging system 1.

[0142] In a second step S32, second output training data are received, comprising intensity distribution data (15'), where the second output training data are related to the second input training data.

[0143] In a third step S33, a second function is trained based on the second input training data and the second output training data by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data.

[0144] In a fourth step S34, the trained, second function 12 is provided.

[0145] Fig. Figure 6 shows a schematic flowchart of an exemplary embodiment of a computer-implemented third training method M4 according to the invention for providing a trained third function 13, which can be carried out, for example, by a third electronic training system 20 with at least one computing unit.

[0146] In a first step S41, third input training data is received, which includes setting data 10' of an imaging system 1.

[0147] In a second step S42, third output training data are received, comprising load parameters (14'), whereby the third output training data are related to the third input training data.

[0148] In a third step S43, a third function is trained based on the third input training data and the third output training data by generating third output data by applying the third function to the third input training data and adjusting parameters of the third function based on a comparison of the third output data with the third output training data.

[0149] In a fourth step S44, the trained, third function 13 is provided.

[0150] Fig. Figure 7 shows a schematic setup of an embodiment of a trained first function 11. In this embodiment, the trained first function 11 can have at least one encoder 21 and one decoder 22.

[0151] In particular, the intensity distribution data 15, shown here, for example, as a graph in which the radiation intensity of a two-dimensional XY area is plotted on a Z-axis, can serve as input data for the encoder, whereby variable-length intensity distribution data 15 can be encoded into a fixed-length sequence. Preferably, the setting data 10 and the load parameter 14 already have a fixed-length sequence and can therefore serve directly as input data for the decoder 22. The decoder 22 is then configured to generate the load distribution data 16 based on the input data, which are shown here, for example, as a graph in which the load intensity of a two-dimensional XY area is plotted on a Z-axis.

[0152] Fig. Figure 8 shows a schematic diagram of an embodiment of a trained second function 12. In this embodiment, the trained second function 12 can have at least one decoder 22. Preferably, the setting data 10 already have a fixed-length sequence and can therefore serve directly as input data for the decoder 22. The decoder 22 is then configured to generate the intensity distribution data 15 based on the input data, which are shown here, for example, as a graph in which the radiation intensity of a two-dimensional XY surface is plotted on a Z-axis.

[0153] Fig. Figure 9 shows an embodiment of an artificial neural network, ANN, 100 for a function or for a trained function, in particular for a trained first function 11 and / or a trained second function 12 and / or for a trained third function 13. Alternative terms for “artificial neural network” are “neural network”, “artificial neural network” or “neural network”.

[0154] The artificial neural network 100 comprises nodes 120, ..., 132 and edges 140, ..., 142, where each edge 140, ..., 142 is a directed connection from a first node 120, ..., 132 to a second node 120, ..., 132. Generally, the first node 120, ..., 132 and the second node 120, ..., 132 are distinct nodes; it is also possible for the first node 120, ..., 132 and the second node 120, ..., 132 to be identical. Fig. For example, edge 140 is a directed connection from node 120 to node 123, and edge 142 is a directed connection from node 130 to node 132. An edge 140, ..., 142 from a first node 120, ..., 132 to a second node 120, ..., 132 is also called an "incoming edge" for the second node 120, ..., 132 and an "outgoing edge" for the first node 120, ..., 132.

[0155] In this embodiment, the nodes 120, ..., 132 of the artificial neural network 100 can be arranged in layers 110, ..., 113, wherein the layers can have an intrinsic order introduced by the edges 140, ..., 142 between the nodes 120, ..., 132. In particular, the edges 140, ..., 142 can exist only between adjacent layers of nodes. In the illustrated embodiment, there is an input layer 110 comprising only nodes 120, ..., 122 without incoming edges, an output layer 113 comprising only nodes 131, 132 without outgoing edges, and hidden layers 111, 112 between the input layer 110 and the output layer 113. In general, the number of hidden layers 111, 112 can be chosen arbitrarily. The number of nodes 120, ..., 122 within the input layer 110 usually refers to the number of input values ​​of the neural network, and the number of nodes 131, 132 within the output layer 113 usually refers to the number of output values ​​of the neural network.

[0156] In particular, each node 120, ..., 132 of neural network 100 can be assigned a (real) number as its value. Here, x(n)i denotes the value of the i-th node 120, ..., 132 of the n-th layer 110, ..., 113. The values ​​of the nodes 120, ..., 122 of the input layer 110 correspond to the input values ​​of neural network 100, and the values ​​of the nodes 131, 132 of the output layer 113 correspond to the output value of neural network 100. Furthermore, each edge 140, ..., 142 can have a weight, which is a real number; in particular, the weight is a real number within the interval [-1, 1] or within the interval [0, 1]. Here, w(m,n)i,j denotes the weight of the edge between the i-th node 120, ..., 132 of the m-th layer 110, ..., 113 and the j-th node 120, ..., 132 of the n-th layer 110, ..., 113. Furthermore, the abbreviation w(n)i,j is defined for the weight w(n,n+1)i,j.

[0157] To calculate the output values ​​of neural network 100, the input values ​​are propagated through the neural network. Specifically, the values ​​of nodes 120, ..., 132 of the (n+1)-th layer 110, ..., 113 can be calculated based on the values ​​of nodes 120, ..., 132 of the n-th layer 110, ..., 113 by xj(n+1)=f(∑ixi(n)⋅wi,j(n)).

[0158] The function f is a transfer function (another term is "activation function"). Well-known transfer functions include step functions, sigmoid functions (e.g., the logistic function, the generalized logistic function, the hyperbolic tangent, the arctangent function, the error function, the smoothstep function), and rectifier functions. The transfer function is primarily used for normalization.

[0159] In particular, the values ​​are propagated layer by layer through the neural network, with the values ​​of the input layer 110 being given by the input of the neural network 100, with the values ​​of the first hidden layer 111 being able to be calculated on the basis of the values ​​of the input layer 110 of the neural network, with the values ​​of the second hidden layer 112 being able to be calculated on the basis of the values ​​of the first hidden layer 111, and so on.

[0160] To determine the values ​​w(m,n)i,j for the edges, neural network 100 must be trained with training data. The training data includes, in particular, training input data and training output data (denoted as ti). In a training step, neural network 100 is applied to the training input data to generate computed output data. Specifically, the training data and the computed output data comprise a number of values ​​corresponding to the number of nodes in the output layer.

[0161] In particular, a comparison between the calculated output data and the training data is used to recursively adjust the weights within the neural network (backpropagation algorithm). Specifically, the weights are changed according to w'i,j(n)=wi,j(n)−γ⋅δj(n)⋅xi(n), where γ is a learning rate, and the numbers δ(n)j can be calculated recursively as δj(n)=(∑kδk(n+1)⋅wj,k(n+1))⋅f'(∑ixi(n)⋅wi,j(n)) based on δ (n+1) j, if the (n+1)th layer is not the output layer, and δj(n)=(xk(n+1)−tj(n+1))⋅f'(∑ixi(n)⋅wi,j(n)) if the (n+1)th layer is the output layer 113, where f' is the first derivative of the activation function and y(n+1)j is the comparison training value for the j-th node of the output layer 113.

[0162] Fig.Figure 10 shows an embodiment of a convolutional neural network 200 for the function or trained function. In the illustrated embodiment, the convolutional neural network 200 comprises an input layer 210, a convolution layer 211, a pooling layer 212, a fully connected layer 213, and an output layer 214. Alternatively, the convolutional neural network 200 can also include multiple convolution layers 211, multiple pooling layers 212, and multiple fully connected layers 213, as well as other types of layers. The order of the layers can be chosen arbitrarily; typically, fully connected layers 213 are used as the last layers before the output layer 214.

[0163] In particular, in a convolutional neural network 200, the nodes 220, ..., 224 of a layer 210, ..., 214 can be viewed as a d-dimensional matrix or as a d-dimensional image. Specifically, in the two-dimensional case, the value of the node 220, ..., 224 indexed by i and j in the nth layer 210, ..., 214 can be denoted as x(n) [i,j]. However, the arrangement of the nodes 220, ..., 224 of a layer 210, ..., 214 has no effect on the computations performed in the convolutional neural network 200 as such, since these are determined solely by the structure and the weights of the edges.

[0164] In particular, a convolution layer 211 is characterized by the structure and weights of the incoming edges that form a convolution operation based on a certain number of kernels. Specifically, the structure and weights of the incoming edges are chosen such that the values ​​x (n)k the node 221 of the folding layer 211 as a folding x (n) k = K k * x (n-1) based on the values ​​x (n-1) The node 220 of the preceding layer 210 is calculated, where the convolution in the two-dimensional case is defined as xk(n)[i,j]=(Kk*x(n−1))[i,j]=∑i,∑j, Kk[i',j']⋅x(n−1)[i−i',j−j'].

[0165] Here, the k-th kernel is K kA d-dimensional matrix (in this case, a two-dimensional matrix) that is typically small compared to the number of nodes 220, ..., 224 (e.g., a 3x3 matrix or a 5x5 matrix). This means, in particular, that the weights of the incoming edges are not independent but are chosen to yield the convolution equation. Specifically, for a kernel that is a 3x3 matrix, there are only 9 independent weights (each entry of the kernel matrix corresponds to an independent weight), regardless of the number of nodes 220, ..., 224 in the respective layer 210, ..., 214. Specifically, for a convolution layer 211, the number of nodes 221 in the convolution layer is equal to the number of nodes 220 in the preceding layer 210 multiplied by the number of kernels.

[0166] If the nodes 220 of the preceding layer 210 are arranged as a d-dimensional matrix, the use of multiple kernels can be interpreted as adding another dimension (referred to as the "depth" dimension), so that the nodes 221 of the convolution layer 221 are arranged as a (d+1)-dimensional matrix. If the nodes 220 of the preceding layer 210 are already arranged as a (d+1)-dimensional matrix with one depth dimension, the use of multiple kernels can be interpreted as extending along the depth dimension, so that the nodes 221 of the convolution layer 221 are also arranged as a (d+1)-dimensional matrix, with the size of the (d+1)-dimensional matrix being larger with respect to the depth dimension by a factor of the number of kernels than in the preceding layer 210.

[0167] The advantage of using convolutional layers 211 is that a spatially local correlation of the input data can be exploited by enforcing a local connectivity pattern between nodes of neighboring layers, in particular by connecting each node only to a small area of ​​the nodes of the preceding layer.

[0168] In the illustrated embodiment, the input layer 210 comprises 36 nodes 220 arranged as a two-dimensional 6x6 matrix. The convolution layer 211 comprises 72 nodes 221 arranged as two two-dimensional 6x6 matrices, each of which is the result of convolution of the values ​​of the input layer with a kernel. Equivalently, the nodes 221 of the convolution layer 211 can be interpreted as a three-dimensional 6x6x2 matrix, the last dimension being the depth dimension.

[0169] A pooling layer 212 can be characterized by the structure and weights of the incoming edges and the activation function of its nodes 222, which form a pooling operation based on a nonlinear pooling function f. In the two-dimensional case, for example, the values ​​x (n) Node 222 of pooling layer 212 based on the values ​​x (n-1) The node 221 of the preceding layer 211 is calculated as follows: x(n)[i,j]=f(x(n−1)[id1, jd2],…,x(n−1)[id1+d1−1, jd2+d2−1])

[0170] In other words, by using a pooling layer 212, the number of nodes 221, 222 can be reduced by replacing a number d1,d2 of neighboring nodes 221 in the preceding layer 211 with a single node 222, which is calculated as a function of the values ​​of the number of neighboring nodes in the pooling layer. The pooling function f can, in particular, be the max function, the average, or the L2 norm. Specifically, in a pooling layer 212, the weights of the incoming edges are fixed and are not changed by the training.

[0171] The advantage of using a pooling layer 212 is that the number of nodes 221, 222 and the number of parameters are reduced. This leads to a reduction in the computational effort in the network and to control overfitting.

[0172] In the illustrated embodiment, the pooling layer 212 is a max-pooling, in which four adjacent nodes are replaced by only one node, where the value is the maximum of the values ​​of the four adjacent nodes. The max-pooling is applied to each d-dimensional matrix of the preceding layer; in this embodiment, the max-pooling is applied to each of the two two-dimensional matrices, thereby reducing the number of nodes from 72 to 18.

[0173] A fully connected layer 213 can be characterized in that a majority, in particular all edges, are present between the nodes 222 of the preceding layer 212 and the nodes 223 of the fully connected layer 213, and wherein the weight of each of the edges can be individually adjusted.

[0174] In this embodiment, the nodes 222 of the preceding layer 212 of the fully connected layer 213 are represented both as two-dimensional matrices and additionally as non-connected nodes (represented as a row of nodes, with the number of nodes reduced for clarity). In this embodiment, the number of nodes 223 in the fully connected layer 213 is equal to the number of nodes 222 in the preceding layer 212. Alternatively, the number of nodes 222 and 223 can also be different.

[0175] Furthermore, in this embodiment, the values ​​of the nodes 224 of the output layer 214 are determined by applying the softmax function to the values ​​of the nodes 223 of the preceding layer 213. Due to the application of the softmax function, the sum of the values ​​of all nodes 224 of the output layer is 1, and all values ​​of all nodes 224 of the output layer are real numbers between 0 and 1. In particular, when using the convolutional neural network 200 to categorize input data, the values ​​of the output layer can be interpreted as the probability that the input data falls into one of the various categories.

[0176] A convolutional neural network 200 can also contain a ReLU layer (acronym for "rectified linear units"). Specifically, the number of nodes and the structure of the nodes in a ReLU layer correspond to the number of nodes and the structure of the nodes in the preceding layer. In particular, the value of each node in the ReLU layer is calculated by applying a rectification function to the value of the corresponding node in the preceding layer. Examples of rectification functions are f(x) = max(0,x), the hyperbolic tangential function, or the sigmoid function.

[0177] Convolutional neural networks 200 can be trained, in particular, using the backpropagation algorithm. To prevent overfitting, regularization methods can be used, e.g., omitting nodes 220, ..., 224, stochastic pooling, using artificial data, weight reduction based on L1 or L2 norms, or max-norm constraints.

[0178] Overall, the examples show how machine learning methods can be used to estimate an X-ray exposure distribution.

[0179] An inhomogeneous X-ray beam leads to an inhomogeneous dose distribution. The resulting estimates of the skin (organ, etc.) dose are incorrect, as the dose is locally higher / lower than determined with the DAP chamber.

[0180] An important aspect of one embodiment of the invention is the prediction of an inhomogeneous beam (dose distribution) using a machine learning model, in particular a trained function that has been trained on measured exposure and detector intensity distributions. An inhomogeneous beam results in an inhomogeneous exposure distribution on the beam and an inhomogeneous intensity distribution at the detector.

[0181] An important aspect of an embodiment of the invention is therefore to use the intensity distribution of the detector of a particular system to “personalize” a general load distribution model.

[0182] An important aspect of one embodiment of the invention is to combine two machine learning models, in particular the trained first function and the trained second function. The trained first function can predict the dose distribution in the beam based on the intensity distribution of the detector (no or homogeneous object in the beam), the tube and collimator settings, and the DAP (dose area product). This function can be trained using dosimetry measurements (e.g., with films) or simulations (e.g., Monte Carlo simulations) of multiple systems and detector images, as well as tube and collimator settings.

[0183] The trained, second function predicts the system-specific detector intensity distribution (no or homogeneous object in the beam) based on the tube and collimator settings. This model can be trained using data (detector images as well as tube and collimator settings) collected, for example, during annual service measurements.

[0184] During application, for example during a medical procedure with a patient in the beam, the two models, i.e. the trained first function and the trained second function, can be combined, which allows a prediction of the dose distribution in the beam based on the settings of the tube and collimator.

[0185] This allows for a more accurate calculation of the X-ray dose. In particular, it ensures that the local peak dose (skin dose) is not underestimated.

[0186] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included.

Claims

[1] Computer-implemented method (M1) for providing data of a radiation exposure distribution (16) from a radiation source (4) of an imaging system (1), comprising the steps: - Receiving setting data (10) of the imaging system (1); - Receiving a radiation exposure parameter (14) associated with the setting data (10); - Receiving data from the setting data (10) associated with an intensity distribution (15) of the radiation (6) at a radiation detector (3) of the imaging system (1); - Applying a trained first function (11) to the setting data (10), the load parameter (14) and the intensity distribution data (15) as first input data of the trained first function (11), whereby the load distribution data (16) are generated as first output data of the trained first function (11); - Providing the load distribution data (16), where the intensity distribution data (15) are provided by applying a trained second function (12) to the setting data (10) as the second input data of the trained second function (12), generating the intensity distribution data (15) as the second output data of the trained second function (12). [2] Computer-implemented method (M1) according to claim 1, wherein the trained second function (12) is provided by a computer-implemented second training method (M3), comprising the steps: - Receiving second input training data, comprising setting data (10') of the imaging system (1); - Receiving second output training data, comprising intensity distribution data (15'), wherein the second output training data are related to the second input training data; - Training a second function based on the second input training data and the second output training data by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data; - Provide the trained, second function (12). [3] Computer-implemented method (M1) according to one of the preceding claims, wherein the load parameter (14) is provided by applying a trained third function (13) to the setting data (10) as the third input data of the trained third function (13), wherein the load parameter (14) is generated as the third output data of the trained third function (13). [4] Computer-implemented method (M1) according to claim 3, wherein the trained third function (13) is provided by a computer-implemented third training method (M4) comprising the steps: - Receiving third input training data, including setting data (10') of the imaging system (1), - Receiving third output training data, comprising load parameters (14'), wherein the third output training data are related to the third input training data; - Training a third function based on the third input training data and the third output training data by generating third output data by applying the third function to the third input training data and adjusting parameters of the third function based on a comparison of the third output data with the third output training data; - Provide the trained third function (13). [5] Computer-implemented method (M1) according to any of the preceding claims, wherein the load parameter (14) is a dose area product or a parameter dependent on the dose area product. [6] Computer-implemented method (M1) according to one of the preceding claims, wherein the intensity distribution data (15) is a two-dimensional X-ray projection image. [7] Computer-implemented method (M1) according to one of the preceding claims, wherein the load distribution data (16) are a two-dimensional image of the load distribution (16) at an interventional reference point (9) of the imaging system (1). [8] Computer-implemented method (M1) according to one of the preceding claims, wherein the setting data (10) of the imaging system (1) comprise at least operating parameters of the radiation source (4) and / or a collimator of the imaging system (1). [9] Computer-implemented method (M1) according to any of the preceding claims, wherein the imaging system (1) is an X-ray-based imaging system and the radiation source (4) is an X-ray radiation source. [10] Computer-implemented training procedure (M2) for providing a trained first function (11) that generates data of a load distribution (16) of radiation (6) from a radiation source (4) of an imaging system (1), and a trained second function (12) that generates data of an intensity distribution (15) of radiation (6) from a radiation source (4) of an imaging system (1), comprising the steps: - Receiving initial input training data, comprehensive - Imaging system settings (10') - radiation exposure parameters (14') associated with the setting data (10'), and - data relating to the setting data (10') of an intensity distribution (15') of the radiation (6) at a radiation detector (3) of the imaging system (1); - Receiving initial baseline training data, comprising load distribution data (16'), wherein the initial baseline training data are related to the initial input training data; - Training a first function based on the first input training data and the first output training data by generating first output data by applying the first function to the first input training data and adjusting parameters of the first function based on a comparison of the first output data with the first output training data; - Provide the trained, first function (11); - Receiving second input training data, comprising setting data (10') of the imaging system (1); - Receiving second output training data, comprising intensity distribution data (15'), wherein the second output training data are related to the second input training data; - Training a second function based on the second input training data and the second output training data by generating second output data by applying the second function to the second input training data and adjusting parameters of the second function based on a comparison of the second output data with the second output training data; - Provide the trained, second function (12). [11] Computer-implemented training method (M2) according to claim 10, wherein the initial training data are provided by dosimetric measurements and / or dosimetric simulations. [12] Data processing device (2) comprising at least one computing unit (7) adapted to perform a computer-implemented method (M1) according to any one of claims 1 to 9. [13] Computer program product (8) comprising instructions which, when the computer program product (8) is executed on a data processing device (2), cause the data processing device (2) to perform the method (M1) according to any one of claims 1 to 9. [14] Imaging system (1) comprising a data processing device (2) according to claim 12, a radiation source (4) for generating radiation (6), and a radiation detector (3) for detecting parts of the ionizing radiation (6), in particular parts of the radiation (6) that have passed through an object (5) to be imaged.

Images