Method for performing an inspection of a stack system

Abstract

The invention relates to a method for performing an inspection of a stacking system. A computer-implemented method for performing an inspection of a stacking system (STACK-SYS) comprises the step of receiving (REC) input data (IN). The method further comprises a step of applying a first function (M1, M1.1, M1.2, M1.3) to (APP-1) the input data, in which intermediate output data (IO1, IO2, IO3) is generated. The method further comprises the step of performing (CHECK) a check based on the intermediate output data (IO1, IO2, IO3). If the check is positive, the method comprises the following steps: providing (PROV-1) the intermediate output data (IO1, IO2, IO3) at least as part of the intermediate input data to the second function (M2); applying a second function (M2) to the (APP-2) intermediate input data, where output data is generated; and providing (PROV-2) the output data.

Description

Technical Field

[0001] The present invention relates to a method for performing an inspection of a stacking system, a stacking system, a system for performing an inspection of a stacking system, a computer program product, and a computer-readable storage medium. Background Art

[0002] Stacking systems are well known in the art. Stacking systems include at least a first function and a second function. Wherein, typically, at least one of the functions includes a trained function. The first function is applied to input data to generate intermediate output data. The intermediate output data is used as input data for a second function, which generates additional intermediate output data or final output data for a third function.

[0003] As such, the first function may be a more general function that can be used to prepare input data for a more specific task. The second function may be an application specific function configured for a specific task in a specific environment.

[0004] In particular, with respect to training functions, this allows the use of a more general first function that can be applied in several use cases. Furthermore, the first function can be trained with a large amount of less specific data. The second function can then be trained on a smaller amount of data specific to the use case of each application. For example, the first function can be configured to segment organs in medical image data. The second function can then be specialized to answer a specific clinical question related to one of the segmented organs.

[0005] However, if the intermediate output is incorrect, this will result in the final output of the second function being incorrect. It is often not obvious to the user where and when the error occurred. This can cause problems in particular in a medical context, where the reliability of the systems used is a particularly critical factor. Summary of the invention

[0006] It is an object of the present invention to provide a method for enhancing the reliability of such a stacking system.

[0007] This object is achieved by a method for performing an inspection of a stacking system, a stacking system, a system for performing an inspection of a stacking system, a computer program product and a computer-readable storage medium according to the invention. Advantageous features and further developments are listed in the following description.

[0008] In the following, the solution according to the invention is described with respect to the claimed system and with respect to the claimed method. Features, advantages or alternative embodiments herein may be assigned to other claimed objects and vice versa. In other words, the claims of the system may be improved with features described or claimed in the context of the method. In this case, the functional features of the method are implemented by the target unit of the system.

[0009] Furthermore, in the following, the solution according to the invention is described with respect to a method and system for performing an inspection of a stacking system and with respect to a method and system for providing a training function. Features, advantages or alternative embodiments herein may be assigned to other claimed objects, and vice versa. In other words, the claims for the method and system for providing a training function may be improved with features described or claimed in the context of a method and system for performing an inspection of a stacking system, and vice versa.

[0010] In particular, the training function used within the method and system for performing an inspection of a stacking system may be adapted by the method and system for providing a training function. Furthermore, the input data used within the method and system for performing an inspection of a stacking system may include advantageous features and implementations of the training input data used within the method and system for providing a training function, and vice versa. Furthermore, the output data used within the method and system for performing an inspection of a stacking system may include advantageous features and implementations of the output training data used within the method and system for providing a training function, and vice versa.

[0011] Independent of grammatical term usage, individuals with either male or female identities are included in the term.

[0012] In a first aspect, the present invention relates to a computer-implemented method for performing an inspection of a stacking system. The method comprises the step of receiving input data. Furthermore, the method comprises the step of applying a first function to the input data, wherein intermediate output data are generated. The method comprises the further step of performing the inspection based on the intermediate output data. If the inspection is positive, the method comprises the steps of providing the intermediate output data at least as a part of the intermediate input data to a second function, applying the second function to the intermediate input data, wherein output data are generated and provided.

[0013] The input data particularly comprises medical image data. The medical image data depicts at least a part of a patient. The patient may be a human or an animal or an object. The medical image data particularly comprises a plurality of pixels, wherein each pixel comprises a pixel value. Wherein the first function may be applied to each pixel value.

[0014] The input data is received by the interface. The input data may be provided by a database and / or by a storage medium and / or by a medical imaging device such as an X-ray system or a magnetic resonance system or a computer tomography system or an ultrasound system.

[0015] By applying the first function to the input data, intermediate output data is generated. The intermediate output data may also include image data. Alternatively, the intermediate output data may include any other data, such as vector data or single values ​​or coordinates associated with pixels of the input medical image data.

[0016] Based on the intermediate output data, a check is performed. During the check, the intermediate output data is checked and verified to be consistent with the rules or prior knowledge, respectively. In particular, during the check, the intermediate output data is compared with the rules and / or prior knowledge about the expected output data. The check may be performed automatically. Alternatively or additionally, the check may be performed manually based on user input.

[0017] In particular, the checking may comprise applying a training function to the intermediate output data. The output of the training function indicates whether the checking is positive or negative.

[0018] A positive result of the check means that the intermediate output data satisfies the rules and / or is consistent with the prior knowledge and therefore appears to be correct. A negative result of the check means that the intermediate output data at least partially does not satisfy the rules or at least partially does not agree with the prior knowledge.

[0019] If the check is positive, the intermediate output data is provided to the second function at least as part of the intermediate input data. In particular, the intermediate input data may include further data, such as input data, in addition to the intermediate output data. In particular, the intermediate input data may include intermediate output data of the third function. The second function is applied to the intermediate input data. Thus, output data is generated. Output data is provided. In particular, the output data may be provided to a user via a user interface as final output data, and / or to a database. Alternatively or additionally, the output data may be provided as second intermediate output data to a third function applied to the output data, thereby generating further output data.

[0020] The invention is not limited to only one first function and one second function. The stacking system may include more than one first function and / or more than one second function. In particular, the stacking system may include a further function that follows the first function and the second function and that is applied to the output data of at least one of the first function and / or the second function in the event of a positive check.

[0021] The method has the advantage that each intermediate output data is checked to be correct in the sense of checking whether it is consistent with rules and / or prior knowledge and / or user knowledge. Based on this, it can be avoided that erroneous intermediate output data are further processed by further functions of the stacking system, which may lead to seriously erroneous final output data. The user is informed of any problems detected during the check at least by a warning message. Therefore, this helps to improve the interpretability of the stacking system. Therefore, the reliability of such a stacking system can be improved, which reliability is a key factor in particular in a medical environment.

[0022] According to an aspect of the invention, if the check is negative, the method comprises the step of providing a warning message.

[0023] If the check is negative, meaning that the intermediate output data appear to be erroneous, a warning message is provided. The warning message may in particular be provided to the user via a user interface. The user interface may in particular be a monitor and / or a loudspeaker and / or a light. The warning message may be configured to stop further processing of the intermediate output data. Alternatively, the warning message may be just information provided to the user that a problem occurred during the check.

[0024] Even if the check is negative and a warning message is optionally provided, the intermediate output data may be provided to the second function and the second function may be applied to the intermediate output data as described above for the case of a positive check. Thus, the generated output data may then be provided as described above.

[0025] Advantageously, the user is informed of problems that arise during the check. Even if the intermediate output data is used as input for the second function, the user knows that he has to check the final output more carefully since it is not safe that everything worked correctly.

[0026] According to an aspect of the invention, the first function comprises a first training function and / or the second function comprises a second training function.

[0027] Typically, the training function mimics cognitive functions that humans associate with other human minds. In particular, through training based on training data, the training function is able to adapt to new situations and is able to detect and infer patterns.

[0028] 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. In addition, representation learning (an alternative term is "feature learning") can be used. In particular, the parameters of the training function can be iteratively adjusted through several steps of training.

[0029] In particular, the training function may include a neural network, a support vector machine, 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 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.

[0030] Each of the training functions, in particular the first training function and the second training function, are trained independently of each other.

[0031] This means that the first training input data and the first training output data are used to train the first function. For training, the first function is applied to the first training input data, wherein first estimated output data are generated. The first estimated output data are compared with the first training output data, and at least one parameter of the first training function is adjusted so that when the first training function is applied again to the first training input data, the first estimated output data determined based on the adjusted first training function are more similar to the first training output data. The first training input data has the same characteristics as the input data.

[0032] The second training function is trained based on the second training input data and the second training output data. The second training input data may include the first training output data. The second training function is trained similarly to the first training function based on the corresponding training data.

[0033] The inventors have realized that the stacking system is particularly beneficial if training functions are used. Like this, different functions can be trained for different use cases based on different and more specific or less specific training data. Like this, for example, a first training function can be trained to be more general than a second training function. This means that the second training function can be more application specific, while the first training function is more like a pre-processing step for the input data of the second training function.

[0034] According to a further aspect of the invention, providing output data comprises at least one of the following steps: providing the output data as final output data to a user and / or a database and / or an imaging device, or providing the output data as second intermediate input data to a third function after performing the examination.

[0035] If the output data is final output data of the stacking system, the final output data may be provided to a user via a user interface, and / or to a database, and / or to an imaging device.

[0036] The user interface may be in particular a monitor that can display the final output data. The user may be in particular a medical professional, such as a doctor and / or a medical assistant. Alternatively, the user may be a patient.

[0037] The database may in particular be a radiology information system (abbreviation: RIS) and / or a picture archiving and communication system (abbreviation: PACS) and / or a hospital information system (abbreviation: HIS) and / or a laboratory information system (abbreviation: LIS) etc. The final output data may in particular be incorporated into an electronic medical patient record which is stored in one of the above-mentioned databases and corresponds to the patient to whom the input data belongs. In this case, the final output data may in particular include a diagnosis and / or a treatment proposal etc.

[0038] The imaging device may in particular be a medical imaging device, such as an X-ray system and / or a magnetic resonance system and / or a computed tomography system and / or an ultrasound system. In this case, the final output data may in particular comprise at least one imaging parameter, the at least one imaging parameter being configured to control the imaging device when acquiring a new image. The imaging parameter may be configured to control image acquisition by means of the medical imaging device.

[0039] If the output data is only the second intermediate output data, a further check is performed based on the second intermediate output data, which may be configured as described above. If the check is positive, the second intermediate output data may be provided to a third function at least as a part of the second intermediate input data. The third function may include a third training function. The second intermediate input data may also include input data of the stacking system.

[0040] If the check is negative, a warning message is optionally provided.As described above, also in this case, the second intermediate output data may optionally be provided to a third function.

[0041] Advantageously, the output data is configured such that it can be used additionally.For this purpose, the output data is provided in a corresponding manner.

[0042] According to another aspect of the present invention, if the output data is provided as final output data, the method further comprises at least one of the following steps: acquiring medical image data based on the output data using a medical imaging system; automatically including the output data in a patient record; and providing a diagnosis based on the output data.

[0043] If the medical image data is acquired based on the final output data, the final output data defines at least one acquisition parameter. The medical imaging system may, for example, include at least one of the following: an X-ray system, a computed tomography system, a magnetic resonance system, an ultrasound system. The final output data is then provided to the medical imaging system so that the medical imaging system can acquire the medical image data based on the final output data.

[0044] Alternatively or additionally, the final output data may be automatically included in a patient record. The patient record is in particular an electronic patient record. The patient record relates to the patient. The input data relates to the same patient as the patient record.

[0045] Alternatively or additionally, a diagnosis may be provided based on the final output data. The diagnosis may be included in a patient record as described above. Alternatively or additionally, the diagnosis may be provided to a user via a user interface, in particular via a monitor. The diagnosis may include information about the disease that the patient to which the input data relates is most likely to have. Alternatively or additionally, the diagnosis does not necessarily need to relate to a disease or anomaly. In particular, the diagnosis may also describe physical facts of the patient, for example, the diagnosis may describe the state of the patient such as the shape of the spine, etc.

[0046] The inventors have realised that the final output data may be used additionally.

[0047] According to a further aspect of the invention, the checking is rule-based.

[0048] Therein, at least one rule that the intermediate output data must satisfy is predefined. The rule may be based on a priori knowledge about the expected output data. Alternatively, the rule may be learned based on a plurality of intermediate output data. In other words, the rule may be an expected value based on a large number of samples of the values ​​of the intermediate output data. The rule may define within which range around the expected value the intermediate output data satisfies the rule, and when the intermediate output data is out of range, the rule is therefore not satisfied.

[0049] If the first function comprises, for example, the segmentation of organs and / or bones in a medical image, the rule may, for example, specify that the number of segmented kidneys is two. Other rules may, for example, specify the expected range of distances between different vertebrae of the spine. There may be a large number of such rules defining the expected range for the intermediate output data.

[0050] The inventors have realised that rule-based checking itself can prove whether the intermediate output data meets expectations / rules which can be based on prior knowledge and can define typical ranges for the intermediate output data. In this way, deviations in the intermediate output data can be detected.

[0051] According to a further aspect of the invention, performing the check includes the steps of providing intermediate output data to a user and receiving user input based on the provided intermediate output data, wherein the user input indicates whether the check is positive or negative.

[0052] The intermediate output may be provided to a user via a user interface. The user interface may include at least a monitor. The intermediate output data may then be displayed via the monitor.

[0053] User input may also be received via a user interface. For this purpose, the user interface may include at least one of the following: a keyboard, a computer mouse, a touch screen, a touch pad, a microphone for voice control, etc. The user input may indicate whether the intermediate output data is correct. If the intermediate output data is correct, the check is positive. If the intermediate output data is incorrect, the check is negative.

[0054] The inventors have realized that including a manual component in the check can be beneficial. For example, in particular in the case where there is a deviation between the intermediate output data and the standard / prior knowledge, it can be checked whether the deviation is correct. For example, some people only have one kidney. In the example mentioned above, when the first function includes a segmentation algorithm, the user can check whether it is correct to segment only one kidney because the patient only has one kidney, or whether it is incorrect to segment only one kidney and the first function failed to segment the second kidney. The inventors have realized that including the user in the check helps to determine erroneous intermediate output data to avoid that the final output data is therefore incorrect.

[0055] According to a further aspect of the invention, if the rule-based check is negative, the intermediate output data is provided to the user.

[0056] In other words, the check comprises two steps. In the first step, the rule-based check described above is performed based on the intermediate output data. If the rule-based check is positive, the intermediate output data is provided to the second function at least as a part of the intermediate input data. If the rule-based check is negative, the second step of the check is then performed. The second step comprises a manual, user-based check as described above. In this case, the user can check via the user interface whether the rule-based check is correct, or whether the rule-based check fails, for example, because the intermediate output data includes an exception to the rule. If the user finds that the rule-based check is already correct, the check is negative. If the user finds that the rule-based check fails, the check is positive.

[0057] Exemplarily, the input data comprises medical image data depicting a patient having one kidney. The first function comprises an organ segmentation function. The first function segments only one kidney. Because the corresponding rule indicates that humans have two kidneys, the rule-based check is negative. Therefore, the segmented medical image data is provided to the user. The user can check whether there is an unsegmented additional kidney, or whether the depicted patient indeed has only one kidney. If the patient has only one kidney, the user can indicate via user input that the first function is correct. In this case, the check is positive.

[0058] The inventors have realized that only selected situations should be presented to the user to avoid overloading the user with work. Advantageously, only those situations where errors may occur are presented to the user. In this case, the user is an example of double checking for unclear situations. In this way, it is possible to avoid that deviations from the standard are correct and are classified as errors and the stacking system is stopped, and on the other hand, it is possible to avoid that the user has to check all each intermediate output data.

[0059] According to another aspect of the invention, the user input includes one of the following: user agreement with the intermediate output data, indicating an affirmative check; user disagreement with the intermediate output data, indicating a negative check; correction of the intermediate output data, wherein the corrected intermediate output data replaces the intermediate output data during a subsequent process, indicating an affirmative check based on the corrected intermediate output data.

[0060] If the user agrees with the intermediate output data, the user indicates that the intermediate output data is correct. In this case, the check is positive.

[0061] If the user disagrees with the intermediate output data, the user indicates that the intermediate output data is wrong. In this case, the check is negative.

[0062] As a third option, the user can correct the intermediate output data. For example, the user can manually correct the wrong segmentation. The corrected intermediate output data replaces the intermediate output data in the subsequent process. Then, it is assumed that the check is positive. The corrected intermediate output data is then provided to the second function at least as part of the intermediate input data.

[0063] The inventors have realized that user input can be used as a check. The inventors have further realized that it can be helpful if the user can correct the intermediate output data so that the stacking system can be further executed with the corrected data. In this way, subsequent errors in the final output data can be avoided while the advantages of the stacking system can be used.

[0064] According to a further aspect of the invention, the input data comprises medical image data, in particular medical radiology image data.

[0065] The medical image data in particular depicts at least a part of a patient. The patient may be a human or an animal. In particular, the medical image data depicts the internal structure of the patient.

[0066] The medical radiology image data may in particular comprise X-ray images and / or computed tomography images and / or magnetic resonance images.

[0067] The inventors have realised that it is beneficial to apply the stacking system to medical image data. In other words, the stacking system can be used to process medical image data. The inventors have realised that a check can be performed on an intermediate output produced by applying the first function to the medical image data.

[0068] According to a further aspect of the invention, the input data comprises medical image data of a spine. The intermediate output data comprises at least one of: the position of at least one vertebra of the spine; a segmentation of the spine; a baseline describing the shape of the spine.

[0069] In other words, the stacking system is configured to analyze the patient's spine. Then, the first function can be configured to determine the position of at least one vertebra of the spine. Alternatively or additionally, the first function can be configured to segment the spine, in particular the vertebrae of the spine. Alternatively or additionally, the first function is configured to determine a baseline of the spine, which baseline describes the shape of the spine.

[0070] The inventors have realised that the described method may be advantageously used to analyse the spine of a patient.

[0071] In a second aspect, the invention relates to a stacking system comprising at least one first function; at least one check gate; and at least one second function. Wherein the at least one first function is configured to be applied to input data, wherein intermediate output data is generated. Wherein the at least one check gate is configured to perform a check on the intermediate output data. And, if the check is positive, the second function is configured to be applied to the intermediate output data, thereby generating output data.

[0072] The stacking system is particularly configured to be used in the method described above.

[0073] According to an aspect of the invention, the checking is rule-based and / or based on user input.

[0074] In particular, the configuration checks are as described above.

[0075] In a third aspect, the present invention relates to a system for performing an inspection of a stacking system, the system comprising an interface and a computing unit. The interface and the computing unit are configured to perform the following steps:

[0076] - receiving input data;

[0077] - applying the first function to the input data,

[0078] wherein, generating intermediate output data;

[0079] - perform checks based on intermediate output data,

[0080] If the check is positive, then

[0081] - providing the intermediate output data to a second function as at least a part of the intermediate input data;

[0082] - applying a second function to the intermediate input data, wherein output data is generated;

[0083] - provide output data, or

[0084] If the check is negative, then

[0085] - Provides a warning message.

[0086] In particular, the system can be configured to perform the previously described method for performing an inspection of a stacking system. The system is configured to perform the method and its aspects through an interface and a computing unit configured to perform the corresponding method steps. In particular, the interface can include one or more sub-interfaces. In particular, the computing unit can include one or more computing sub-units.

[0087] In another aspect, the present invention relates to a training system for providing a stacking system, wherein the stacking system includes a first function and a second function. The first function includes a first training function and / or the second function includes a second training function. The training system includes a training interface and a training computing unit, and the training interface and the training computing unit are configured to perform the following steps:

[0088] If the first function includes a first training function:

[0089] - receiving first input training data and first output training data;

[0090] - training a first function based on first input training data and first output training data,

[0091] If the second function includes a second training function:

[0092] - receiving second input training data and second output training data,

[0093] Wherein, the first output training data is the same as the second input training data;

[0094] - training a second function based on second input training data and second output training data;

[0095] -Providing a stacking system comprising a first function and a second function.

[0096] In a fourth aspect, the invention relates to a computer program product having a computer program and a computer readable medium. A predominantly software-based implementation has the advantage that even previously used systems can be easily upgraded by software updates to work in the described manner. In addition to the computer program, such a computer program product may optionally include additional components such as documentation and / or additional components, and hardware components such as a hardware key (dongle etc.) for using the software.

[0097] In a further aspect, the invention relates to a computer program product comprising program elements directly loadable into a storage unit of a first providing system, which, when the program elements are executed by the system, causes the system to perform a method according to the claimed method and its aspects.

[0098] In a fifth aspect, the invention relates to a computer-readable storage medium comprising program elements readable and executable by a system to perform the claimed method and aspects thereof when the program elements are executed by the system.

[0099] In a further aspect, the invention relates to a computer-readable storage medium comprising a stacking system as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] Other objects and features of the present invention will become apparent from the following detailed description considered in conjunction with the accompanying drawings. It should be understood, however, that the accompanying drawings are designed for purposes of illustration only and not as a definition of the limits of the present invention.

[0101] Figure 1 A schematic flow chart of an exemplary stacking system known in the prior art is shown.

[0102] Figure 2 shows a schematic flow chart of a first specific embodiment of a method for performing an inspection of a stacking system,

[0103] Figure 3 a schematic flow chart showing a first embodiment of the steps for performing a check,

[0104] Figure 4 a schematic flow chart showing a second embodiment of the steps for performing a check,

[0105] Figure 5 a schematic flow chart showing a third embodiment of the steps for performing a check,

[0106] Figure 6 shows a schematic flow chart of a stacking system including an inspection door,

[0107] Figure 7 A schematic flow chart showing the processing of data by applying a method for performing an inspection of a stacking system, and

[0108] Figure 8 A system for performing an inspection of a stacked system is shown. DETAILED DESCRIPTION

[0109] Figure 1 A schematic flow chart of an exemplary stacking system known in the prior art is shown.

[0110] The stacking system comprises a plurality of functions M1.1, M1.2, M1.N, M2.1, M2.2, M2.N, M3. Input data IN is provided to at least the first layer functions M1.1, M1.2, M1.N. The first layer functions M1.1, M1.2, M1.N are applied to the input data, so that each of the first layer functions M1.1, M1.2, M1.N generates intermediate output data. The intermediate output data can be used as intermediate input data for functions M2.1, M2.2, M2.N of another layer. In particular, a combination of two or more intermediate output data can be provided as intermediate input data to a second layer function M2.1, M2.2, M2.N. In particular, the intermediate input data of at least one of the second layer functions M2.1, M2.2, M2.N can additionally include input data IN.

[0111] The stacked system may comprise one or more layers of such functions M1.1, ..., M2.N.

[0112] The intermediate output data of the second layer functions M2.1, M2.2, M2.N are provided as intermediate input data to a further function M3, which is applied to the intermediate output data of the second layer functions M2.1, M2.2, M2.N. The intermediate input data of the further function M3 may additionally include input data IN. M3 generates final output data OUT.

[0113] One or more of the functions M1.1, ..., M2.N, M3 may comprise a training function. The general principle of a training function is described above.

[0114] Figure 2 A schematic flow chart of a first specific embodiment of a method for carrying out an inspection of a stacking system is shown.

[0115] In the step of receiving REC input data IN, input data IN is received from a medical system or a database. In particular, the input data IN may include medical image data, in particular medical radiology image data. The medical image data includes an image of at least a part of a patient. The patient may be a human, an animal or an object. The medical radiology image data may in particular include an X-ray image and / or a tomography image and / or a magnetic resonance image and / or a computer tomography image and / or an angiography image, etc.

[0116] In the step of applying the first function to the APP-1 input data IN, the first function is applied to the input data IN, thereby generating intermediate output data.

[0117] Based on the intermediate output data, a check is performed in the step of performing a CHECK check. During the check, it is checked whether the intermediate output data is correct. Figures 3 to 5 A more detailed description of different examples of how to implement this check is described in .

[0118] When the intermediate output data appears to be correct, the check is positive, otherwise the check is negative.

[0119] If the check is positive, the intermediate output data are provided to the second function in a step of providing PROV-1 intermediate output data at least as part of the intermediate input data.The intermediate input data may also particularly comprise input data and intermediate output data of other functions comprised by the stacking system.

[0120] The second function is applied to the intermediate output data, wherein the output data is generated.

[0121] In the step of providing PROV-2 output data, the output data is provided for further use.

[0122] In particular, the output data may be final output data OUT as the final output of the stacking system. In this case, the output data is no longer used within the stacking system for generating further data. In this case, the final output data may be provided in the step of providing PROV-1 the output data to a user and / or a database and / or an imaging device.

[0123] If output data are provided to a user, the output data may in particular be displayed on a monitor.

[0124] If the output data is provided to a database, the output data can in particular be automatically incorporated into a patient record relating to the patient depicted for the corresponding input data. Alternatively or additionally, the database can be used to automatically generate a diagnosis based on the provided output data. In other words, the output data provided to the database can be used in further applications to further analyze and / or use the output data, to generate further knowledge, and / or to store the output data in a structured, understandable manner, and / or to generate parameters that can be used for further data generation, etc.

[0125] In particular, the output data may be provided to a medical imaging system that may use the output data to generate new medical image data. In this case, the output data may include imaging parameters. In X-ray imaging, the imaging parameters may be, for example, acquisition time and / or tube voltage and / or tube current and / or position of the patient, etc.

[0126] Alternatively, the generated output data is not final. In particular, the output data may be second intermediate output data. The second intermediate output data may be provided to the third function at least as a part of the second input data. In this case, the steps of checking described above are repeated for the output generated by the third function.

[0127] The stacking system may include a plurality of such functions. Each of such functions may receive intermediate input data, which may include, for example, input data IN and / or intermediate output data of one or more other functions. One of these functions may be configured to generate the final output of the stacking system.

[0128] In particular, the first function may comprise a first training function. In particular, the second function may comprise a second training function. General examples of training functions are provided above.

[0129] If the check is negative, a warning message is optionally provided in the step of providing a PROV-3 warning message. The warning message may in particular be provided to the user. The warning message may be provided via a monitor and / or via a loudspeaker and / or via a flashing light etc. The warning message informs the user that the stacking system does not appear to be functioning properly based on the input data IN.

[0130] In particular, even if the check is negative, the intermediate output data may be provided to the second function as described above. The only difference may be that the user is informed of potential problems that have been detected during the check.

[0131] Figure 3 A schematic flow chart of a first embodiment showing the steps of performing a CHECK inspection.

[0132] In this embodiment, the checking includes the step of performing a CHECK-R rule-based check.

[0133] The rule-based check may be based on at least one hard-coded rule. The rule depends on what is displayed by the input data. For example, the rule may predefine the number of kidneys that a patient typically has (two). And / or the rule may predefine the typical distances between individual vertebrae of the spine and the corresponding possible deviations, etc.

[0134] The rule-based check may in particular comprise a fourth training function. The fourth training function may be applied to the intermediate output data, wherein information is generated informing how well the intermediate output data fits the expectation. For this purpose, the fourth training function may be trained based on a large number of intermediate output data, the large number of intermediate output data being generated by applying the first function to a large number of medical image data depicting the same patient part as the input data. The fourth training function may then compare the intermediate output data to the large number of intermediate output data and may therefore determine whether the intermediate output data meets the expectation or whether there are some significant deviations from the expectation.

[0135] If there is a deviation from the rules and / or expectations, the check is assumed to be negative. Otherwise, the check is assumed to be positive.

[0136] Figure 4 A schematic flow chart of a second embodiment of the steps for performing a CHECK inspection is shown.

[0137] The step of performing the CHECK check includes the step of providing the intermediate output data PROV-U to the user, and the step of receiving REC-U the user input based on the provided intermediate output data. The user input indicates whether the check is positive or negative. This can also be referred to as a user-based check.

[0138] In the step of providing PROV-U to the user with intermediate output data, the intermediate output data can be provided to the user in particular via a user interface such as a monitor. Like this, the user can view the intermediate output data and check whether the intermediate output data seems to be correct.

[0139] The user input may be provided by the user via a user interface. For this purpose, the user interface may include a computer mouse and / or a touch pad and / or a touch screen and / or a keyboard and / or a speaker for voice control, etc.

[0140] Via the user input, the user can indicate whether the intermediate output seems correct according to his perception, and as such, he can indicate whether the check is positive or negative.

[0141] Figure 5A schematic flow chart of a third embodiment showing the steps of performing a CHECK inspection.

[0142] According to this embodiment, the steps of performing the CHECK inspection are as follows: Figure 3 and Figure 4 A combination of implementation methods.

[0143] In a first step, a rule-based check is performed.

[0144] If the rule-based check is negative, then Figure 4 As described for the user-based check, the intermediate output data is provided to the user. The user-based check described is then performed.

[0145] If the rule-based check is positive, then the completed check is assumed to be positive and the method is as described in Figure 1 Continue as described.

[0146] Figure 6 A schematic flow chart of a stacking system STACK-SYS including a check gate CG is shown.

[0147] The stacking system comprises at least one first function M1 applied to input data IN. The input data IN may be configured as described above. After being applied to the input data IN, the first function M1 generates intermediate output data. The intermediate output data must pass a check gate CG. The check gate CG may perform a rule-based check and / or a user-based check. The check may be configured as described above. In the depicted embodiment, the rule-based check C1 is combined with a user-based check UI. In a first step, it is evaluated whether the intermediate output data conforms to predefined rules and / or conforms to expectations based on various similar data. This is called a rule-based check C1 and is performed according to Figure 3 If the rule-based check C1 is positive, the intermediate output data is provided to the second function at least as part of the intermediate input data. If the rule-based check C1 is negative, the intermediate output data is additionally provided to the user via the user interface. The user can then perform the Figure 4 In this step, the user can evaluate whether the negative result of the rule-based check C1 is correct. If the user also indicates that the check is negative, a warning message is provided and further data processing is stopped. If the user indicates that the check should be positive, the intermediate output data is provided to the second function at least as part of the intermediate input data.

[0148] The first function M1 and the check gate CG form a first complex N1. The second function is part of a second complex N2. The second complex may also include further check gates.

[0149] The stacking system STACK-SYS may comprise a further complex N3 and thus further functions. One of the functions provides the final output FO provided by the stacking system STACK-SYS as described above.

[0150] Figure 7 A schematic flow chart of the processing of data by applying a method for carrying out an inspection of a stacked system STACKED-SYS is shown.

[0151] The input data IN includes an X-ray image of a human spine. The input data IN is provided to three first functions M1.1, M1.2, M1.3, each of which generates corresponding intermediate output data IO1, IO2, IO3. The first first function M1.1 provides a segmentation of the spine. The second first function M1.2 provides a contour of the spine. The third first function M1.3 provides the location of specific vertebrae of the spine. The intermediate output data IO1, IO2, IO3 of each first function M1.1, M1.2, M1.3 is provided to a corresponding inspection gate CG1.1, CG1.2, CG1.3, each inspection gate including as described in accordance with Figure 6 The described rule-based checks C1.1, C1.2, C1.3 and the user-based check UI. If at least one of the user-based check or the rule-based check is positive, the intermediate output data IO1, IO2, IO3 are provided to the second function M2 at least as part of the intermediate input data. The second function M2 provides a final output FO. The final output FO includes the shape of the spine in the X-ray image.

[0152] Figure 8 A system SYS for carrying out an inspection of a stack system STACK-SYS is shown.

[0153] The system SYS shown is configured to carry out the method according to the invention for carrying out an inspection of a stack system STACK-SYS. The system SYS comprises an interface SYS.IF, a calculation unit SYS.CU and a memory unit SYS.MU.

[0154] The system SYS may in particular be a computer, a microcontroller or an integrated circuit. Alternatively, the system SYS may be a real network or a virtual network of computers (the technical term for a real network is a "cluster" and the technical term for a virtual network is a "cloud"). The system SYS may also be designed as a virtual system executed on a computer, a real network of computers or a virtual network of computers (the technical term is "virtualization").

[0155] 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 have hardware elements or software elements, such as a microprocessor or a so-called FPGA (short for "field programmable gate array"). The memory unit SYS.MU can be implemented as a non-permanent working memory (random access memory, RAM for short) or as a permanent mass storage device (hard disk, USB stick, SD card, solid state drive).

[0156] The interface SYS.IF may in particular comprise a plurality of sub-interfaces for carrying out different steps of the respective method. In other words, the interface SYS.IF may also be understood as a plurality of interfaces SYS.IF. The computing unit SYS.CU may in particular comprise a plurality of sub-computing units for carrying out different steps of the respective method. In other words, the computing unit SYS.CU may also be understood as a plurality of computing units SYS.CU.

[0157] In the absence of explicit description, the individual embodiments or their various aspects and features can be combined or exchanged with each other without limiting or expanding the scope of the described invention, as long as such combination or exchange makes sense and is within the meaning of the present invention. Where applicable, the advantages described with respect to one embodiment of the present invention are also beneficial to other embodiments of the present invention.

Claims

1. A computer-implemented method for performing an inspection of a stack system (STACK-SYS), comprising the following steps: - Receive (REC) input data (IN); - applying (APP-1) a first function (M1, M1.1, M1.2, M1.3) to said input data, Among them, intermediate output data (IO1, IO2, IO3) is generated; - performing (CHECK) a check based on said intermediate output data (IO1, IO2, IO3), If the check is positive, then - providing (PROV-1) said intermediate output data (IO1, IO2, IO3) to a second function (M2) at least as a part of the intermediate input data; - applying (APP-2) said second function (M2) to said intermediate input data, wherein output data are generated; -Provide (PROV-2) said output data.

2. The method according to claim 1, in, If the check is negative, the method comprises the following steps: - Provide (PROV-3) warning message.

3. The method according to one of the preceding claims, in, The first function (M1, M1.1, M1.2, M1.3) comprises a first training function, and / or Wherein, the second function (M2) comprises a second training function.

4. The method according to one of the preceding claims, in, Providing the output data includes: - providing said output data as final output data (FO) to a user and / or a database and / or an imaging device, or - providing said output data as second intermediate input data to a third function after performing the check.

5. The method according to claim 4, If the output data is provided as final output data, the method further comprises at least one of the following steps: - acquiring medical image data based on the final output data (FO) using a medical imaging system; - automatically including said final output data (FO) in the patient record; - providing a diagnosis based on said final output data (FO).

6. The method according to one of the preceding claims, in, The checking is rule-based.

7. The method according to one of the preceding claims, in, Performing the inspection includes the following steps: - providing (PROV-U) the intermediate output data (IO1, IO2, IO3) to a user; - receiving (REC-U) user input based on the provided intermediate output data (IO1, IO2, IO3), Wherein, the user input indicates whether the check is positive or negative.

8. The method according to claim 7 in combination with claim 6, in, If the rule-based check is negative, the intermediate output data (IO1, IO2, IO3) are provided to the user.

9. The method according to claim 7 or 8, in, The user input includes one of the following: - said user agrees with said intermediate output data (IO1, IO2, IO3), indicating a positive check; - the user disagrees with the intermediate output data (IO1, IO2, IO3), indicating a negative check; - correction of the intermediate output data (IO1, IO2, IO3), wherein the corrected intermediate output data replaces the intermediate output data (IO1, IO2, IO3) during a subsequent process, indicating a positive check based on the corrected intermediate output data.

10. The method according to one of the preceding claims, in, The input data (IN) comprises medical image data, in particular medical radiology image data.

11. The method according to claim 10, in, The input data (IN) includes medical image data of the spine, The intermediate output data (IO1, IO2, IO3) includes at least one of the following: - the position of at least one vertebra of the spine; - segmentation of the spine; - A baseline describing the shape of the spine.

12. A stacking system (STACK-SYS), comprising - at least one first function (M1, M1.1, M1.2, M1.3); - at least one inspection door (CG); and - second function (M2), in, The at least one first function (M1) is configured to be applied to input data (IN), wherein intermediate output data (IO1, IO2, IO3) are generated, wherein the check gate (CG) is configured to perform a check of the intermediate output data (IO1, IO2, IO3), wherein, if the check is positive, the second function (M2) is configured to be applied to the intermediate output data, The second function (M2) is configured to be applied to the intermediate output data (IO1, IO2, IO3) to generate output data.

13. A stacking system (STACK-SYS) according to claim 12, in, The checking is rule-based and / or based on user input.

14. A system (SYS) for performing a check of a stack system (STACK-SYS), the system comprising an interface (SYS.IF) and a calculation unit (SYS.CU), wherein: The interface (SYS.IF) and the computing unit (SYS.CU) are configured to perform the following steps: - receiving input data (IN); - applying a first function (M1, M1.1, M1.2, M1.3) to said input data (IN), Among them, intermediate output data (IO1, IO2, IO3) is generated; - performing a check based on said intermediate output data (IO1, IO2, IO3), If the check is positive, then - providing said intermediate output data (IO1, IO2, IO3) to a second function (M2) at least as a part of the intermediate input data; - applying said second function (M2) to said intermediate input data, wherein output data are generated; - provide said output data, or If the check is negative, then - Provides a warning message.

15. A computer program product comprising program elements which can be directly loaded into a storage unit of a system (SYS), the computer program product causing the system (SYS) to perform a method according to any one of claims 1 to 11 when the program elements are executed by the system (SYS).

16. A computer-readable storage medium comprising program elements that can be read and executed by a system (SYS) to perform the method according to one of claims 1 to 11 when the program elements are executed by the system (SYS).

Images