CN110568471A - Method for determining threshold values of energy bands, computing unit and medical imaging device - Google Patents
Method for determining threshold values of energy bands, computing unit and medical imaging device Download PDFInfo
- Publication number
- CN110568471A CN110568471A CN201910491851.3A CN201910491851A CN110568471A CN 110568471 A CN110568471 A CN 110568471A CN 201910491851 A CN201910491851 A CN 201910491851A CN 110568471 A CN110568471 A CN 110568471A
- Authority
- CN
- China
- Prior art keywords
- ray detector
- additional information
- imaging device
- examination object
- direct
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 77
- 238000002059 diagnostic imaging Methods 0.000 title claims abstract description 31
- 238000006243 chemical reaction Methods 0.000 claims abstract description 46
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 44
- 238000004364 calculation method Methods 0.000 claims abstract description 19
- 238000012360 testing method Methods 0.000 claims description 25
- 238000004590 computer program Methods 0.000 claims description 24
- 238000012549 training Methods 0.000 claims description 14
- 238000009826 distribution Methods 0.000 claims description 7
- 239000007943 implant Substances 0.000 claims description 7
- 238000010801 machine learning Methods 0.000 claims description 7
- 229910052751 metal Inorganic materials 0.000 claims description 7
- 239000002184 metal Substances 0.000 claims description 7
- 230000002308 calcification Effects 0.000 claims description 3
- 239000000203 mixture Substances 0.000 claims description 3
- 238000003771 laboratory diagnosis Methods 0.000 claims description 2
- 239000003550 marker Substances 0.000 claims description 2
- 238000013528 artificial neural network Methods 0.000 description 22
- 238000012545 processing Methods 0.000 description 14
- 238000002591 computed tomography Methods 0.000 description 11
- 238000003384 imaging method Methods 0.000 description 9
- 230000015654 memory Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 7
- 230000005855 radiation Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 210000001519 tissue Anatomy 0.000 description 6
- 230000008901 benefit Effects 0.000 description 5
- 238000003860 storage Methods 0.000 description 5
- 208000000913 Kidney Calculi Diseases 0.000 description 4
- 206010029148 Nephrolithiasis Diseases 0.000 description 4
- 210000003484 anatomy Anatomy 0.000 description 4
- 210000000746 body region Anatomy 0.000 description 4
- ZCYVEMRRCGMTRW-UHFFFAOYSA-N 7553-56-2 Chemical compound [I] ZCYVEMRRCGMTRW-UHFFFAOYSA-N 0.000 description 3
- 210000001015 abdomen Anatomy 0.000 description 3
- 210000004027 cell Anatomy 0.000 description 3
- 210000000038 chest Anatomy 0.000 description 3
- 238000004891 communication Methods 0.000 description 3
- 239000002872 contrast media Substances 0.000 description 3
- 238000010191 image analysis Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 229910052740 iodine Inorganic materials 0.000 description 3
- 239000011630 iodine Substances 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000002093 peripheral effect Effects 0.000 description 3
- 229910004613 CdTe Inorganic materials 0.000 description 2
- 208000008589 Obesity Diseases 0.000 description 2
- 230000004913 activation Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 229920005994 diacetyl cellulose Polymers 0.000 description 2
- 201000010099 disease Diseases 0.000 description 2
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000036541 health Effects 0.000 description 2
- 238000003703 image analysis method Methods 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000027939 micturition Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 235000020824 obesity Nutrition 0.000 description 2
- 238000012015 optical character recognition Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- OYPRJOBELJOOCE-UHFFFAOYSA-N Calcium Chemical compound [Ca] OYPRJOBELJOOCE-UHFFFAOYSA-N 0.000 description 1
- 208000002513 Flank pain Diseases 0.000 description 1
- 229910001218 Gallium arsenide Inorganic materials 0.000 description 1
- 206010027457 Metastases to liver Diseases 0.000 description 1
- 241001465754 Metazoa Species 0.000 description 1
- 208000002193 Pain Diseases 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000002238 attenuated effect Effects 0.000 description 1
- 238000013529 biological neural network Methods 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 229910052791 calcium Inorganic materials 0.000 description 1
- 239000011575 calcium Substances 0.000 description 1
- 230000001201 calcium accumulation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 239000002800 charge carrier Substances 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000003745 diagnosis Methods 0.000 description 1
- 239000003814 drug Substances 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 230000003862 health status Effects 0.000 description 1
- 210000001981 hip bone Anatomy 0.000 description 1
- 210000004394 hip joint Anatomy 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 208000017169 kidney disease Diseases 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012886 linear function Methods 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 229910052710 silicon Inorganic materials 0.000 description 1
- 239000010703 silicon Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 239000000758 substrate Substances 0.000 description 1
- 230000003936 working memory Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/42—Arrangements for detecting radiation specially adapted for radiation diagnosis
- A61B6/4208—Arrangements for detecting radiation specially adapted for radiation diagnosis characterised by using a particular type of detector
- A61B6/4241—Arrangements for detecting radiation specially adapted for radiation diagnosis characterised by using a particular type of detector using energy resolving detectors, e.g. photon counting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/161—Applications in the field of nuclear medicine, e.g. in vivo counting
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/161—Applications in the field of nuclear medicine, e.g. in vivo counting
- G01T1/163—Whole body counters
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/24—Measuring radiation intensity with semiconductor detectors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/24—Measuring radiation intensity with semiconductor detectors
- G01T1/247—Detector read-out circuitry
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/16—Measuring radiation intensity
- G01T1/24—Measuring radiation intensity with semiconductor detectors
- G01T1/249—Measuring radiation intensity with semiconductor detectors specially adapted for use in SPECT or PET
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01T—MEASUREMENT OF NUCLEAR OR X-RADIATION
- G01T1/00—Measuring X-radiation, gamma radiation, corpuscular radiation, or cosmic radiation
- G01T1/29—Measurement performed on radiation beams, e.g. position or section of the beam; Measurement of spatial distribution of radiation
- G01T1/2914—Measurement of spatial distribution of radiation
- G01T1/2985—In depth localisation, e.g. using positron emitters; Tomographic imaging (longitudinal and transverse section imaging; apparatus for radiation diagnosis sequentially in different planes, steroscopic radiation diagnosis)
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H40/00—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices
- G16H40/60—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices
- G16H40/63—ICT specially adapted for the management or administration of healthcare resources or facilities; ICT specially adapted for the management or operation of medical equipment or devices for the operation of medical equipment or devices for local operation
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/02—Arrangements for diagnosis sequentially in different planes; Stereoscopic radiation diagnosis
- A61B6/03—Computed tomography [CT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/48—Diagnostic techniques
- A61B6/488—Diagnostic techniques involving pre-scan acquisition
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/52—Devices using data or image processing specially adapted for radiation diagnosis
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B6/00—Apparatus or devices for radiation diagnosis; Apparatus or devices for radiation diagnosis combined with radiation therapy equipment
- A61B6/54—Control of apparatus or devices for radiation diagnosis
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- General Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- High Energy & Nuclear Physics (AREA)
- Theoretical Computer Science (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Public Health (AREA)
- Software Systems (AREA)
- Biophysics (AREA)
- Optics & Photonics (AREA)
- Radiology & Medical Imaging (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Animal Behavior & Ethology (AREA)
- Pathology (AREA)
- Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Veterinary Medicine (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
the invention relates to a method of determining threshold values of energy bands, a calculation unit and a medical imaging device. Method for determining a threshold value of at least one energy band of a direct conversion X-ray detector of a medical imaging device comprising the steps of: acquiring a scout image of an examination object and/or at least one additional information about the examination object, determining at least one feature of the examination object from the scout image and/or the additional information, determining the threshold value based on the determined feature using a machine-learned trained algorithm.
Description
Technical Field
The invention relates to a method for determining a threshold value of at least one energy band of a direct conversion X-ray detector of a medical imaging device.
Background
A new technology in the field of X-ray imaging, in particular computed tomography, is the direct conversion type or direct conversion type detector. Here, instead of using a scintillator which first converts X-ray photons into visible light, a semiconductor is used which directly detects the X-ray photons. Advantageously, a direct conversion type detector is capable of resolving the energy of incident photons in addition to the position, for example in spectral computed tomography. Here, instead of determining the energy individually for each photon, the photons are associated with a specific energy band as a function of the energy and the events are counted separately for each pixel for each energy band. Typically four or less energy bands, for example. In this way, the amount of data is reduced to a practical extent. The threshold value of the energy band is determined beforehand on the part of the manufacturer and the electronics in the detector module are set accordingly. For this purpose, it can be provided that either a fixed, unchanging threshold value can be preset or can be selected on the part of the user depending on or together with the desired scanning protocol, for example for head or chest recordings. In order to optimize the medically relevant information content of the acquired image records, the threshold values should be adjusted in a manner that matches the initial suspicion, matches the intended diagnosis or matches the interrogation. In addition, the threshold value should also be determined taking into account the presence of foreign bodies, such as, for example, kidney stones, implants, in particular of metal, or contrast agents, in order to take into account the corresponding imaging properties. However, this is not practically properly implemented in protocol planning for users such as radiologists or medical operators in hospitals or similar medical facilities. There is no automatic method, so that hitherto the measurement has been carried out with the aid of preset and possibly suboptimal thresholds.
Disclosure of Invention
In contrast, it is an object of the present invention to provide an alternative mechanism which allows a reliable, automatic and reproducible determination of the threshold value of at least one energy band of a direct-conversion X-ray detector. In particular, it is an object of the invention to take into account the individual anatomical features and/or health conditions of the patient.
This object is achieved by a method for determining a threshold value of at least one energy band of a direct-conversion X-ray detector of a medical imaging device, a corresponding computing unit and medical imaging device, a corresponding computer program and a corresponding computer-readable data carrier. Preferred and/or alternative, advantageous embodiment variants are the subject of the following examples.
the solution according to the invention of said object is described next with respect to the claimed method and with respect to the claimed device. In this case, the mentioned features, advantages or alternative embodiments also apply to the other claimed subject matter and vice versa. In other words, the features described or claimed in relation to particular embodiments (e.g. as directed to a method) can also be exploited to advantage in connection with the apparatus. The corresponding functional features of the method are formed by the corresponding specific modules or units.
The present invention relates in a first aspect to a method for determining a threshold value of at least one energy band of a direct conversion X-ray detector of a medical imaging device.
For the purposes of the present invention, direct conversion X-ray detectors are also referred to as counting detectors or counting detectors for energy discrimination. The core element of these detectors is a direct conversion type material (e.g. CdTe, CdZTe ═ CZT, Si or GaAs) that directly converts incident X-ray radiation, i.e. individual X-ray quanta, into electrical signals. A charge pulse of a height corresponding to the energy of an X-ray quantum is recorded as a counting event if it exceeds a defined threshold. If a plurality of thresholds are provided for each pixel element of the X-ray detector, the thresholds form different energy bands with respectively defined boundaries in which events can be counted as a function of the associated quantum energy. Direct conversion type detectors are characterized by high quantum efficiency. In embodiments having multiple thresholds for each pixel element, the thresholds are particularly suitable for multi-energy applications or material separation.
That is, the threshold defines the energy band of the direct conversion type X-ray detector. The energy bands represent a specific range of quantum energies of the X-ray photons. By defining different energy bands by means of a plurality of thresholds, X-ray quanta having different quantum energies in different channels respectively associated with one energy band can be recorded. Typically, four energy bands are provided and three thresholds are set accordingly. The threshold value may be set or adjusted within a certain tolerance range.
According to the invention, at least one threshold value is determined for one energy band or at least one threshold value is determined for at least one energy band, in other words a specific energy value (keV) is determined as a threshold value. It is also possible according to the invention to determine and adjust a plurality of threshold values, for example two threshold values, of the X-ray detector, wherein the remaining threshold values of the X-ray detector are determined by default settings. Alternatively, all thresholds of the X-ray detector, for example three thresholds, can also be determined according to the invention.
The method comprises a plurality of steps.
In a first step of the method according to the invention, a scout image of the examination object and/or at least one additional information about the examination object is acquired.
For the purposes of the present invention, scout images correspond to classical two-dimensional X-ray overlay photographs. Which measures the individual X-ray attenuation distribution through an examination object along a specific projection direction in which the examination object is traversed by X-ray radiation and the X-ray radiation is imaged according to different gray values. The X-ray attenuation is usually used by dose automation devices in computed tomography to determine suitable tube current curves or to perform tube current modulation. For example, it is known to capture scout images of the examination subject in the lateral direction and scout images of the examination subject in the anteroposterior direction before the X-ray images are captured, and to determine the X-ray attenuation distribution for the examination subject in each case from the gray values in the respective directions. Scout images typically image the entire body or large partial areas of the body of the examination subject in one picture, for example the thorax/abdomen. In this connection, the positioning image also includes indications of anatomical features of the examination subject, for example metallic implants or calcium accumulations. During the method according to the invention, a scout image can be recorded, that is to say acquired by means of a medical imaging device. Alternatively, the method according to the invention can acquire scout images by: the generated image files, for example in DICOM format, are loaded into the computing unit from a memory, for example a PACS of a hospital system.
In the context of the present invention, the examination subject is a patient, most of which refers to a human being. In principle, the patient can also be an animal. Therefore, the two terms "examination subject" and "patient" are used synonymously hereinafter. The examination object can alternatively be a plant or an inanimate object, such as a historical relic or the like.
Additional information about the patient is any information describing or characterizing the examination object. The additional information can comprise personal instructions about the examination subject, about its anatomy or also about its disease state, its medical history, its disease course etc. At least one of these specifications is collected as additional information. The at least one additional information can be present, for example, in the form of electronic image files, text files and/or voice files, etc. In particular, a voice file can be input through voice recognition of a user. According to the present invention, one additional information can be collected, but a plurality of additional information can also be collected.
The acquiring can include: image data acquisition is carried out with the aid of a medical imaging device, i.e. an electronic image file is generated, for example an electronic image file of a scout image. The acquiring can also include: the already existing files are loaded into the computing unit. The acquiring can further include: the existing image, text and/or voice files are loaded, for example from a database, a PACS system, a RIS (radiology department information system) or a HIS (hospital information system) or a cloud storage with restricted access.
In a second step, at least one characteristic of the examination object is determined from the scout image and/or the additional information.
In which the scout image and/or at least one additional information is examined or analyzed in order to derive or determine a feature characterizing the examination object. The determination can include image analysis if the scout image or additional information in the form of an image file is evaluated in the determination step. In the context of the present invention, image analysis can include structure recognition, such as structure recognition of anatomical structures, pattern recognition, text recognition, segmentation, texture analysis, etc., to extract features from the image file. The determination also comprises text recognition, in which for example additional information in the form of a text file or a speech file is checked for the presence of keywords or text blocks. The speech file can be converted into a text file beforehand by means of conversion. At least one characteristic is determined according to the invention for characterizing the examination object. In other words, a plurality, i.e. a large number of, for example, three or five, features can also be identified. Preferably, the determination of the at least one characteristic is carried out automatically by means of a computing unit. Alternatively, the determination can be performed semi-automatically, wherein a user input is required, for example an automatically generated suggestion for the determined feature is selected or confirmed. It can also be provided that the specific features can only be determined manually by the user and, for example, entered manually.
In a third step of the method according to the invention, a threshold value is determined with a machine-learned trained algorithm based on the determined features. In other words, a threshold value is automatically derived from the measured features, which are used as input values for a machine-learned trained algorithm.
In the context of the present invention, machine learning includes computer-implemented techniques in which an algorithm is identified based on existing data, patterns, or legitimacy and solutions are derived autonomously by applying the algorithm against unknown new data. A prerequisite for autonomously discovering solutions is a training phase in which machine-learned algorithms are applied to a known, defined and mostly very large database in order to find rules or predictions that achieve a desired output or a desired result. The training can be designed as supervised or unsupervised training, wherein in a first variant the algorithm is provided with value pairs in the form of input values and the correct output values depending therefrom, while in a second variant the algorithm has to autonomously adjust itself on the basis of the input values such that it provides the correct output values.
The machine-learned algorithm is particularly advantageously designed as an artificial neural network. Artificial neural networks follow the construction of biological neural networks, such as the human brain. The artificial neural network preferably comprises a plurality of further layers between the input layer and the output layer, which further layers each comprise at least one node. Each node here corresponds to a processing unit, similar to a biological neuron. Nodes within a layer of the network can be connected to nodes of other layers via directed connections (edges). The connections define data flows within the network. Each node thus represents an operation that is applied to the input data. Further, each node or each connection of nodes has a weight parameter (weight). Via the weight parameter, the influence or importance of the output of the node as input value of the receiving node is defined. In a training phase, which is preferably designed as supervised learning, the artificial neural network "learns" the weight parameters for all nodes or connections from the training data and adjusts these weight parameters until the output layer of the network provides the correct output values.
The described method process is based on the inventor's knowledge on the one hand: by evaluating regularly present and available information, scout images and/or additional information about the examination object, at least one characterizing feature can be derived or determined, preferably automatically. This feature provides information about the individual imaging properties to be expected when performing a targeted image data acquisition by means of the medical imaging device and thus enables an individual adjustment of the threshold values of the direct conversion X-ray detector. The method according to the invention is also based on the following knowledge: the machine-learned trained algorithm establishes, during its training, a fixed relationship between input values, here in the form of features, and output values, for example in the form of threshold values of a direct-conversion X-ray detector corresponding to the features. Accordingly, the determined feature, or advantageously a combination of features or a combination of features comprising at least one feature, is used as an input to a trained algorithm for machine learning. In this respect, the invention advantageously achieves: the threshold of the direct conversion X-ray detector is automatically and individually matched to the examination object for image data acquisition. The image quality or clinical efficacy of the image record can be improved by such matching. The method according to the invention advantageously enables to override the user input.
according to a preferred embodiment of the method according to the invention, the additional information about the examination object is additional information from the group of: a previous image recording of the examination object; an image recording protocol that was applied in a previous image data acquisition on an inspection object; inquiring; referral sheet (Zuweisung); previous test result reports on the test subject or test result reports on laboratory diagnoses of the test subject, etc. All of this information has in common that they each comprise at least one characteristic about the patient. From the previously used image recording protocol, for example, a set tube voltage can be determined, which, for example, allows the obesity of the patient to be inferred. For example, the suggestion of a specific (past) condition of a patient can be derived from the test result report of a laboratory diagnosis, and the referral typically includes an initial suspicion that already led to a targeted image data acquisition. It is clear that all information from the above-mentioned group can be analyzed, even if the features so determined may be at least partially redundant. In practice, however, not all of this information is present or available, so that the additional information considered according to the invention is reduced to the extent that it is available or desired on the user side.
According to a particularly advantageous embodiment of the method according to the invention, at least one additional information is extracted from the electronic medical record. An electronic medical record is a collection of all information available to a patient. Including instructions on the individual as well as information on his health status. Can include image data as well as any other information. In particular, the electronic medical record can comprise a previous image recording of the examination object according to the above-described embodiment; an image recording protocol that was applied in a previous image data acquisition on an inspection object; inquiring; a referral order; previous test result reports on the test subject or test result reports on laboratory diagnoses of the test subject. The electronic medical record thus enables simple access to the scout image or additional information. The electronic medical record can on the one hand be a digital medical record of a medical facility, for example a hospital or a radiology center, and is locally reserved there for each patient in the RIS, HIS and/or PACS. On the other hand, electronic medical records can also be stored on a storage medium, for example a health insurance chip card, and carried around by the patient and provided in preparation for image data acquisition.
According to a preferred embodiment of the method according to the invention, the determined features are anatomical features relating to the examination subject and features from the group of: markers, tissue composition, tissue distribution, width or size of the object under examination, metal implants, calcification or presence of metal implants in the patient, body regions or partial regions of the patient, such as the head or chest, etc. The anatomical features are characterized in that they describe the anatomy of the examination object to such an extent that imaging properties of the individual, which can be taken into account in the image data acquisition by adjusting the threshold values accordingly, are derived therefrom. The anatomical features can thus provide, inter alia, information about the image artifacts to be expected, attenuation characteristics that are characteristic for the patient, etc. The anatomical features can be determined in particular advantageously from the image file comprised by the scout image and/or the additional information by means of image analysis methods known per se.
According to a further embodiment of the method according to the invention, the characteristic is information from the group of the following characteristics: protocol parameters of previous image recording protocols, e.g. tube voltage set at the X-ray sourceParameters indicative of initial suspicion, previous clinical test results and/or test results of laboratory diagnostics on the test subject. The features identified according to this embodiment are likewise characterized in that they describe the anatomy of the examination object to such an extent that imaging properties of the individual, which can be taken into account by adjusting the threshold values accordingly during the image data acquisition, are derived therefrom. The features of this embodiment can be determined, inter alia, by analyzing test result reports, usually text data.
According to a further embodiment of the method according to the invention, the at least one feature is determined by applying a text recognition algorithm to the scout image and/or the additional information. Digital text Recognition, also known as Optical Character Recognition (OCR), is referred to as an algorithm that can recognize by finding patterns, words or text from a raster graphics file, such as an image file of the DICOM standard. With the present invention, the text recognition algorithm can further include: existing text is also checked in terms of content, particularly with respect to the presence of keywords such as "kidney stones" or the like, where the keywords can represent specific features. It can also be provided to use speech recognition algorithms with similar functionality in determining the features.
According to a further embodiment of the method according to the invention, the at least one feature is also determined by means of a machine-learned trained algorithm. That is, in this embodiment, not only the correlation between the extracted features and the threshold values for the direct conversion type X-ray detector is performed by the machine-learned algorithm, but also the measurement of the features itself is performed by the machine-learned algorithm. Particularly preferably, the two steps of the method according to the invention are carried out by means of an algorithm. Particularly suitable for this purpose are algorithms known as deep learning, for example in the form of "convolutional neural networks", also known as volumetric neural networks. In other words, according to this embodiment, a feature extraction is first performed by means of a machine-learned algorithm, and then a so-called classification is performed, wherein the identified features are associated with a defined set of thresholds comprising at least one threshold of the energy band of the X-ray detector. Alternatively to convolutional neural networks, it is also possible to use long-short-term memory (LSTM long short-term memory) networks or Recurrent Neural Networks (RNNs) which have a backward feedback loop within the hidden network layer with respect to what has been mentioned before.
According to another embodiment of the method according to the invention, the determining of the threshold value comprises: selecting a clinical presentation from a plurality of clinical presentations. The clinical representation here represents the overall medical or anatomical state of the patient, which is derived from the entirety of the determined features. Preferably, in this embodiment, the selected clinical presentation is output as an output value by a machine-learned algorithm. Each clinical performance is further associated with a defined set of thresholds, the set comprising at least one threshold of a direct conversion X-ray detector. The association rules can be stored, for example, in a database, which is in data exchange with the computing unit. In this regard, the determination of the threshold value of the at least one energy band of the X-ray detector can comprise: the computing unit reads a set of thresholds corresponding to the output of the algorithm from a database. The method process advantageously achieves: the user is still able to manually change the preselection of the set of threshold values for the automatic determination.
In an alternative embodiment of the invention, the machine learning algorithm directly outputs the set of thresholds corresponding to the set of features as an output value. The set of thresholds is particularly preferably capable of being transmitted directly to a direct-conversion X-ray detector.
According to a further embodiment of the method according to the invention, the training of the machine-learned algorithm is configured as supervised learning as already explained above and is carried out on the basis of a plurality of pairs of a feature set comprising at least one measured feature and a corresponding associated set of predetermined thresholds comprising at least one threshold of the direct conversion X-ray detector. In this case, error feedback (back-propagation) methods are generally used, which consist in returning the deviation between the actual output value and the expected output value known from the training data to the artificial neural network in order to adjust the respective weighting of the nodes.
In an alternative embodiment, the training of the machine-learned algorithm is designed as a supervised learning and is carried out on the basis of a plurality of pairs of feature sets and associated clinical representations for the examination subject, the feature sets comprising at least one measured feature. The training data used in each case can be based on a plurality of actually performed examinations of the patient or generated manually by simulation.
Another aspect of the invention relates to a calculation unit for determining a threshold value of at least one energy band of a direct conversion X-ray detector of a medical imaging device. The computing unit has means for carrying out the method according to the invention. In a preferred embodiment, the calculation unit is connected to a medical imaging device comprising a direct conversion X-ray detector. Alternatively, the calculation unit is directly connected with the direct-conversion X-ray detector. The calculation unit comprises a control unit which is configured to set a specific threshold value on the direct-conversion X-ray detector.
Advantageously, the calculation unit is integrated into the medical imaging device. Alternatively, the computing unit can also be provided remotely or remotely from the medical imaging device. The computing unit can be designed to carry out the determination of the threshold value, or also the entire method according to the invention, for the medical imaging device or a plurality of devices, for example in a radiology center or hospital comprising a plurality of medical imaging devices, on the basis of the determined characteristics, in particular using machine-learned algorithms.
A third aspect of the invention relates to an X-ray detector in the form of a direct conversion X-ray detector comprising a calculation unit according to the invention.
A fourth aspect of the invention relates to a computer program with a program code for implementing the method according to the invention for determining a threshold value of at least one energy band of a direct conversion X-ray detector, when the computer program is executed on a computer or on a computing unit according to the invention.
A fifth aspect of the invention relates to a computer-readable data carrier (13) with a program code of a computer program for implementing the method according to the invention for determining a threshold value of at least one energy band of a direct conversion X-ray detector when the computer program is executed on a computer, for example a computing unit according to the invention. Advantageously, the step of determining the threshold value can be performed in particular by means of a machine-learned algorithm on a computer, for example in a computing unit of the medical imaging device, on the basis of the determined features.
The implementation of the invention in the form of a computer program or in the form of a computer-readable data carrier comprising program code of a computer program according to the invention provides the following advantages: existing computer systems or computing units can easily be adapted by software upgrades in order to implement the functionality according to the invention.
Alternatively, the computer program can be embodied in the form of a computer program product and have additional units. The additional unit can be configured as hardware, for example as a storage medium on which a computer program and/or a hardware key is stored in order to be able to use the computer program. Alternatively or additionally, the additional unit can be configured as software, for example as a program document or a software key, in order to be able to use the computer program.
A sixth aspect of the invention relates to a medical imaging device comprising a detector of the direct conversion type and a calculation unit according to the invention. A medical imaging apparatus is an imaging apparatus used in medicine. The medical imaging device according to the invention uses X-ray radiation to generate images. In this connection, the medical imaging device is in an alternative embodiment of the invention embodied in the form of a C-arm device, a computed tomography scanner or an X-ray device.
Drawings
The above-described features, characteristics and advantages of the present invention, and the manner and method of attaining them, will become apparent and be better understood by reference to the following description of embodiments, which are set forth in detail in connection with the accompanying drawings. The invention is not limited to these embodiments by the description. In the different figures, identical components are provided with the same reference numerals. The drawings are generally not to scale. The figures show:
Figure 1 shows a schematic view of a method according to the invention according to one embodiment of the invention,
Figure 2 shows a schematic view of a neural network for use in the method according to the invention,
Fig. 3 shows an X-ray detector according to the invention according to an embodiment of the invention, an
Fig. 4 shows a medical imaging device in the form of a computed tomography scanner comprising a calculation unit according to the invention according to a further embodiment of the invention.
Detailed Description
Fig. 1 shows a schematic view of a method according to the invention in one embodiment. In a first step S1, a scout image of the examination object and/or at least one additional information about the examination object is acquired. The acquisition can comprise, in particular in the case of a planned computed tomography examination, an image data acquisition of an X-ray tomogram for which the threshold value is to be determined and adjusted according to the invention. The steps of collecting generally include: a scout image and/or any additional information about the patient is obtained. Additional information is for example an existing clinical image examination (bildstuden) on the patient or a previous single image record, for example an earlier scout image; an image recording protocol that was applied in a previous image data acquisition on an inspection object; inquiring; a referral order; previous test result reports or test result reports for laboratory diagnostics. The acquiring can include: the user enters the additional information manually or by voice via a suitable user interface (see description of fig. 4).
The additional information and/or localization images are preferably present as DICOM objects, in other words they correspond to the DICOM standard, which includes, for example, DICOM images or DICOM structured reports or the like. Other data formats such as pdf, jpeg, rtf, etc., for example.
Preferably, a plurality of different additional information is collected and at least the additional information is considered in the following. It is particularly advantageous to be able to use the further additional information in order to verify or confirm the features derived from the first additional information. This makes the method according to the invention particularly reliable. The additional information can be extracted, provided or called from different data sources, for example the additional information can be called from a central or distributed data store, for example in the form of a patient database, for example from a RIS, HIS and/or PACS or cloud storage. The electronic medical records can be saved on the data memory in a retrievable manner for each patient, so that all available electronic data and documents that can be used as additional information are merged at the (virtual) location. Alternatively, the electronic medical record can also be saved, at least in part, on the patient's personal insurance card.
In a second step S2, at least one characteristic of the examination object is determined from the scout image and/or the at least one additional information. In other words, the electronic data acquired or available about the patient is analyzed here in view of the included features. The characteristic is preferably an anatomical characteristic relating to the examination object, such as the presence of a marker, a specific tissue composition, a specific tissue distribution, the width or size of the examination object, the presence of a metal implant or calcification. The markers or tissue components and/or tissue distributions can specify body regions or body parts of the examination subject, which body parts/body regions should be examined spectroscopically. They can also be used as features characterizing the patient, respectively. The height and width of the patient can provide a prompt for specific recording conditions and thus also enable an inference of the threshold value of the adaptation of the direct conversion X-ray detector. Thus, depending on whether the patient is an obese patient or a lean patient, or depending on whether a child or an adult is to be examined, the patient-specific threshold can be varied, for example.
Furthermore, the feature can be one of the following: protocol parameters of a previous image recording protocol, parameters indicating an initial suspicion, test results of earlier previous clinical tests and/or laboratory diagnoses on the examination subject. A possible protocol parameter is, for example, the X-ray tube voltage used. This X-ray tube voltage can likewise provide an explanation about the obesity of the patient, for example. It was initially suspected, that is to say a specific interrogation, that a body region or a specific examination type, for example an examination with the aid of a contrast agent, can likewise be specified. This applies to the existing condition and/or to clinical conditions which can be deduced from existing (test result) reports.
Particularly advantageously, a plurality of the above-mentioned features are determined in step S2. These features are then processed in the next steps according to the invention. This enables the threshold value to be adapted to the patient particularly precisely.
In step S2, methods for image analysis, for example known algorithms for segmentation, for labeling, for text recognition or structure/pattern/texture recognition, can be used, in particular in the case of determining anatomical features from image data, for example scout images. Furthermore, the method according to the invention also makes use of known methods of text or speech recognition, such as NLP (natural language Processing) algorithms.
In step S3, a threshold or thresholds are determined based on the measured features using a machine-learned trained algorithm. In this step, at least one of the determined characteristics is taken into account. Preferably, the threshold value is determined taking into account a plurality of measured characteristics. Preferably, not only the threshold values, but also a set of threshold values comprising a plurality of threshold values, i.e. at least two threshold values, of at least two energy bands of a direct conversion X-ray detector, which is explained in detail below with reference to fig. 3, are derived in step S3.
In one embodiment of the invention, the machine learning algorithm collects a set of determined features as input data or input. The input data or input is algorithmically converted to a threshold or set of thresholds. In other words, the algorithm outputs a set of thresholds as output values or as output. In step S3, at least one of the determined features is "translated" into a set of thresholds that match the patient. These threshold values can then be transmitted directly, for example via a corresponding control unit, to the direct-conversion detector and set there. No manual intervention by the user is required. In another embodiment, the algorithm derives clinical manifestations from these features. The clinical presentation is then output as an output value. The clinical representation describes a medical or anatomical overall state of the patient, which results taking into account the determined characteristic. Different clinical manifestations are associated with different threshold sets. In this embodiment, step S3 further includes: the clinical performance is associated with the matched threshold set. This is preferably done via access to a database in which the relationships between the plurality of clinical manifestations and the associated threshold sets are maintained. This embodiment enables the user to subsequently manually adjust the threshold sets determined according to the present invention as soon as needed or desired.
Step S2 may be performed in one embodiment of the present invention by means of a machine-learned algorithm like step S3. In this case, steps S2 and S3 are combined. These two steps S2 and S3 can be performed by the same algorithm, for example. In this embodiment, the scout image and/or the at least one additional information act as input values to the algorithm. The determined features are in this case output only as an intermediate result of the hidden layer as input values for at least one subsequent layer.
In an optional step S4, the determined set of thresholds is transmitted to and set at the direct conversion X-ray detector.
In general, features are preferably automatically extracted from current clinical information (e.g., medical records, referrals) or suitable interactive dialogs for manually entering additional information or features and/or scout images, which are used to determine thresholds. These threshold values are transmitted to a detector module of the medical imaging device, for example, a computed tomography scanner, and are subsequently used for a computed tomography scan, i.e., an image data acquisition. For processing the additional information and the scout image, methods from the field of artificial intelligence, such as neural networks, are preferably used. In particular, for evaluating the scout image, an image analysis method is preferably used, preferably an analysis method in natural language is used to evaluate the clinical supplementary information. These methods can either generate the threshold or the threshold set directly as output values or can generate clinical manifestations out of a limited set of possible clinical manifestations, which threshold sets are associated with the possible clinical manifestations via a database, respectively. In this way, a patient-specific optimization of the threshold values of the X-ray detector is achieved.
In a first example, a scout image is acquired along with a referral sheet of the patient. In the scout image, for example, markings such as the lower rib arch and/or hip bone are detected. As a first feature, the "abdomen" of the body part is determined from the scout image. In the referral sheet there are: "undefined pain in the flank, discomfort during urination". The terms "flank pain" and "urination" are recognized by means of text. These terms are associated with the initial suspicion of "kidney disease" as other features, taking into account the first feature "abdomen". From this, the clinical expression "renal calculus" was deduced. For this clinical presentation, a set of thresholds is saved that is best suited to distinguish calcified from non-calcified kidney stones, e.g., 40keV-50keV-80keV (the range around the K-edge of calcium, approximately between 40keV to 50keV, is most contrasted).
In a second example, a referral sheet is collected as additional information. The determination of "liver metastases" was initially suspected as characteristic by means of text recognition. Since such examinations are always performed with the aid of iodine-containing contrast agents, a threshold set is assigned which has at least one low threshold close to the K-edge of iodine (approximately 33keV) in order to achieve the greatest possible contrast in respect of the iodine distribution. If in this example there is also a scout image of the patient, the body width of the patient can thus be determined as a further feature. If the body width of the patient has a value which is less than a predetermined limit value, the patient is then identified as slim or normal weight and the following threshold value sets are associated: 40keV-60keV-80keV (the range between 40keV and 60keV is the most intense). If the body width of the patient has a value greater than a predetermined boundary value, the patient is considered obese and the following set of thresholds is associated: 50keV-70keV-90keV, since the lower energy X-ray photons are more strongly absorbed.
In a third example, an endoprosthesis is determined from the scout image as a characteristic "metal implant", such as an artificial hip joint, which has a high X-ray attenuation. In this case, a set of thresholds is associated or determined which allows metal artifacts to be advantageously suppressed, for example 70keV-100keV-120keV (the higher the X-ray energy, the higher the fraction of transmitted X-ray photons).
Fig. 2 shows an artificial neural network 400, as can be used in the method according to fig. 1. Neural network 400 is responsive to input nodes x for a plurality of applicationsi410 to produce one or more outputs oj. The neural network 400 learns in this embodiment by: the neural network adjusts the weight coefficient w of each node based on training datai(weight). Input node xiPossible input values of 410 can be, for example, anatomical or other features characterizing the patient, which have been extracted beforehand from the scout image and/or the additional information. Alternatively, when the neural network 400 is configured to also perform the feature extraction according to step S2, the input value can be the scout image or the additional information itself. Any other input value can be applied. The neural network 400 weights 420 the input values 410 based on a learning process. The output value 440 of the neural network 400 preferably corresponds to the measured clinical performance or set of thresholds. The output 440 can be via a single or multiple output nodes ojThe process is carried out.
The artificial neural network 400 preferably includes a hidden layer 430 that includes a plurality of nodes hj. Can be provided with a plurality of hidden layers hjnWherein the hidden layer 430 uses the output value of another hidden layer 430 as an input value. The nodes of the hidden layer 430 perform mathematical operations. Node hjCorresponds here to its input value xjAnd a weight coefficient wjOf (2) isA linear function f. At the time of obtaining the input value xjAfter that, node hjPerforming for each input value xjby a weight coefficient wjThe sum of the weighted multiplications, as determined by the function:
hj=f(∑ixi·wij)#
In particular, node hjis formed as a function f of the node activation, for example an S-shaped function or a linear ramp function. Output value hjIs transmitted to one or more output nodes oj. Recalculating each output value hjAs a function of node activation, f:
oj=f(∑ihi·w′ij)#
The neural network 400 shown here is a feed-forward neural network in which all nodes 430 process the output values of the preceding layers in the form of their weighted sum as input values. Of course, other neural network types can also be used according to the invention, for example a feedback network, in which the node hjCan also be its output value.
The neural network 400 is trained by means of supervised learning in order to recognize patterns. The known method process is back propagation, which can be applied to all embodiments of the present invention. The neural network 400 is applied to training input values during training and must produce corresponding a priori known output values. The mean square error between the calculated and expected output values (mean square error- "MSE") is iteratively calculated and the respective weight coefficients 420 are adjusted until the deviation between the calculated and expected output values is less than a predetermined threshold.
Fig. 3 shows a schematic structure of a counting-type or direct conversion-type X-ray detector 9. The X-ray detector of the counting type or the direct conversion type is constituted by a plurality of detector modules 35. The detector module 35 includes a plurality of pixel elements 15, an ASIC27, a direct conversion type material or a section of the direct converter 24, and a coupling between the direct converter 24 and the ASIC27 (e.g., bump bonds 36). ASIC27 is provided inOn the substrate 37 and connected to peripheral electronics 38. The detector module 35 can also have one or more ASICs 27 and one or more subcomponents of the direct converter 24, selected accordingly as required. For large-area X-ray detectors (e.g. 20X 30 cm)2) A plurality of detector modules 35 can be combined (e.g. 11 × 17 modules) and connected via common peripheral electronics 38. For the connection between the ASIC and the peripheral electronics 38, for example, TSV technology (through silicon via) is used in order to ensure the closest possible four-side sequentiality of the modules.
Incident X-ray radiation is converted in a direct converter 24 (e.g., CdTe or CZT) and the generated charge carrier pairs are separated via an electric field generated by a common top electrode 26 and a pixel electrode 25 specific to each pixel element 15. The charge generates a charge pulse in one of the pixel-like pixel electrodes 25 of the ASIC27, the height of which corresponds to the energy of the X-ray quantum and which is recorded as a counting event as soon as it is greater than a defined threshold value. That is, events are counted in the pixel elements 12 of the counting-type X-ray detector 9 by: when the electrical signal generated in the direct converter 24 and corresponding to the incident X-ray quanta is greater than one of a plurality of settable thresholds, a digital memory cell is counted, with a separate memory cell associated with each threshold and the corresponding energy band. The threshold value is determined by means of a discriminator. The threshold value can in principle be preset in a fixed, analog manner, but is usually applied via a digital-to-analog converter (DAC) and can thus be set variably within a range. The threshold can be set either pixel by means of a local discriminator and DAC, or globally for a plurality/all of the pixel elements 15 by means of a global discriminator and DAC. For example, a plurality of global DACs are required for pixel elements 15 having a plurality of thresholds and counters in order to set the thresholds. The direct-conversion X-ray detector 9 shown here has four different threshold values for each pixel element 15, which can be set by means of four global pairs of DACs and discriminators and, depending on this (discriminator threshold), attributes the charge pulse to one or more of the digital memory cells (timers). The X-ray quanta counted in a particular energy region or band can be obtained by differencing the count content of two corresponding counters. In a preferred embodiment, the control unit 22 of the calculation unit 12 according to the invention (see description of fig. 4) intercepts the threshold values determined according to the invention from the calculation unit 12 and transmits them to the pair of DAC discriminators.
Fig. 4 shows a medical imaging device 1 in the form of a computed tomography scanner 1. The computer tomography scanner 1 shown here has an acquisition unit 17 which comprises an X-ray radiation source 8 and an X-ray radiation detector or X-ray detector 9. The X-ray detector 9 is configured as a direct conversion X-ray detector 9 and is configured, for example, as described with reference to fig. 3. Other embodiments are also contemplated. The acquisition unit 17 rotates about the system axis 5 during the acquisition of the X-ray projections, and during the acquisition the X-ray source 8 emits an X-ray beam 2 which passes through the patient 3 and is attenuated there and impinges on the X-ray detector 9.
The patient 3 lies on an examination table 6 while taking X-ray projections. The examination bed 6 is thus connected with the bed base 4, so that said bed base carries the examination bed 6 and the patient 3. The examination table 6 is designed to move the patient 3 in the recording direction through the opening 10 of the acquisition unit 17. The acquisition direction is usually given by the system axis 5 about which the acquisition unit 17 rotates when the X-ray projections are acquired. In this example, the body axis of the patient 3 is identical to the system axis 5. In the case of a helical recording, the table 6 is continuously moved through the opening 10, while the acquisition unit 17 is rotated around the patient 3 and X-ray projections are recorded. The X-ray beam 2 thus impinging on the surface of the patient 3 is helical. The patient 3 may not move, but must not move during the examination.
The computer tomograph 1 has a computing unit 12 in the form of a computer which is connected to a display unit 11, for example for graphically displaying a medical image record, here in the form of a computed tomography photograph or a control menu for the imaging apparatus 1, and to the input unit 7. The display unit 11 can be, for example, an LCD screen, a plasma screen or an OLED screen. Furthermore, it can be a touch-sensitive screen, which is also designed as an input unit 7. Such touch sensitive screens can be integrated into imaging devices or be formed as part of mobile devices. The input unit 7 is, for example, a keyboard, a mouse, a so-called "touch screen" or a microphone for voice input. The input unit 7 can be configured to recognize the movement of the user and to translate into corresponding instructions. By means of the input unit 7, in particular by voice or keyboard, the user can, for example, enter at least one additional information, for example in the form of an inquiry, an initial suspicion about the patient 3, etc.
The computing unit 12 is connected to a rotatable acquisition unit 17 for data exchange. Via the connection 14, on the one hand, control signals for data acquisition or image data acquisition are transmitted from the computer 12 to the acquisition unit 17, in particular via the connection 14, thresholds determined for the different energy bands of the X-ray detector 9 are transmitted to the X-ray detector in order to set the thresholds. On the other hand, the projection data recorded for the patient 3 can be transmitted to the computer 12 for image reconstruction by means of customary reconstruction methods. The connection 14 is realized in a known manner, either wired or wireless.
The computing unit 12 in the form of a computer according to the described embodiment comprises a data processing unit 16. The data processing unit is configured in particular to carry out all the calculation steps that are present with regard to the method according to the invention for the scout image, which has been recorded, for example, by means of the acquisition unit 17, and/or for at least one additional information. The scout image and/or the additional information can also be provided to the data processing unit 16 by another medical imaging device and do not have to be acquired in time immediately before further processing by the data processing unit 16. The scout image and/or the additional information can be supplied to the image data processing unit 16 in a manner known per se, for example via a mobile, per se known computer-readable data carrier, via a PACS, hospital or radiology information system (HIS or RIS), or via the internet, for example from a cloud storage.
The data processing unit 16 comprises means for implementing the method according to the invention, as already described with reference to fig. 1 and 2. The data processing unit 16 therefore comprises a determination unit 21 for determining at least one characteristic of the examination object from the scout image and/or the additional information. Furthermore, the image data processing unit 16 also comprises a determination unit 23 configured to determine a threshold value based on the determined features using a machine-learned trained algorithm. The two units 21 and 23 can be designed as separate processing units, but can also be designed jointly in one unit. The latter is advantageous in particular when only one machine learning algorithm performs the steps of feature determination and threshold determination. In the first case, the units 21 and 23 are at least in data connection, in order to transmit the characteristics determined by the unit 21 to the unit 23 for further processing.
The calculation unit 12 also comprises a control unit 22 via which the calculation unit is in data communication with the direct-conversion X-ray detector 9. The control unit 22 collects the threshold values determined by the data processing unit 16 and transmits them to the DAC discriminator pairs of the X-ray detector 9.
The computing unit 12 can communicate with the mentioned components or units, in particular, via the DICOM standard exchange protocol. Other communication protocols and data formats are contemplated as well.
The computing unit 12 can interact with a computer-readable data carrier 13, in particular for carrying out the method according to the invention by means of a computer program having a program code. Furthermore, the computer program can be stored in a callable manner on a machine-readable data carrier. The machine-readable carrier can be, inter alia, a CD, DVD, blu-ray disc, memory stick or hard disk. The computing unit 12 and thus also its sub-components can be constructed in hardware or in software.
The computation unit 12 is configured, for example, as a so-called FPGA (acronym of "Field Programmable Gate Array" in english) or includes an arithmetic logic unit. The computing unit 12, i.e. the individual or all subcomponents thereof, can alternatively be arranged distributed, for example, the individual computing steps of the method can be carried out in a medical service facility, for example, a central computing center of a hospital or in the cloud. In this case, data and patient protection are taken into account, in particular, during data exchange. The computing unit 12 can furthermore alternatively be designed as a subcomponent of the X-ray detector 9 or as a cloud-based computer, wherein the data exchange with the imaging device 1 and/or the X-ray detector 9 takes place via a secure internet connection. The communication is in a preferred embodiment via the DICOM standard, however other standards or data formats are equally feasible.
In the embodiment shown here, at least one computer program is stored in a memory of the computing unit 12, which computer program executes all method steps of the method according to the invention when the computer program runs on the computer 12. The computer program for performing the method steps of the method according to the invention comprises program code. Furthermore, the computer program can be configured as an executable file and/or stored on a computing system different from the computer 12. The computer tomograph 1 can be designed, for example, such that the computer 12 loads a computer program for carrying out the method according to the invention into its internal working memory via an intranet or via the internet. Whether explicitly stated or not, the various embodiments, individual sub-aspects or features thereof can be combined or interchanged with one another as far as meaningful and in terms of the present invention, without departing from the scope of the present invention. The advantages of the invention described with reference to one embodiment are also applicable to other embodiments as long as they can be reversed, unless otherwise explicitly stated.
Claims (16)
1. A method for determining a threshold value of at least one energy band of a direct conversion X-ray detector (9) of a medical imaging device (1), the method comprising the steps of:
Acquiring (S1) a scout image of an examination object (3) and/or at least one additional information about the examination object,
-determining (S2) at least one characteristic of the examination object from the scout image and/or the additional information,
-determining (S3) the threshold value based on the determined features using a trained algorithm of machine learning (400).
2. the method according to claim 1, wherein the additional information about the examination object is additional information out of the group of: a previous image recording of the examination object; an image recording protocol that was applied in a previous image data acquisition with respect to the examination object; inquiring; a referral order; a previous test result report on the test subject or a test result report on a laboratory diagnosis of the test subject.
3. The method according to claim 1 or 2, wherein at least one of the additional information is extracted (S1) from an electronic medical record.
4. The method according to any of the preceding claims, wherein the features are anatomical features with respect to the examination object and features out of the group of: a marker, a tissue composition, a tissue distribution, a width or size of the examination object, a metal implant or a calcification.
5. The method according to any of the preceding claims, wherein the feature is a feature out of the group of features: protocol parameters of a previous image recording protocol; a parameter indicative of an initial suspicion; previous clinical test results and/or test results of laboratory diagnoses on the test subject.
6. The method according to any of the preceding claims, wherein the at least one feature is determined with a text recognition algorithm applied to the scout image and/or the additional information.
7. The method according to any of the preceding claims, wherein determining (S2) the at least one feature is also performed by means of a trained algorithm of the machine learning (400).
8. The method of any of the above claims, wherein determining the threshold comprises: selecting a clinical presentation from a plurality of clinical presentations, wherein the selected clinical presentation is output by the machine-learned algorithm and each clinical presentation is associated with a defined set of thresholds of the direct-conversion X-ray detector.
9. Method according to one of the preceding claims, wherein the training of the machine-learned algorithm is configured as supervised learning and is performed on the basis of a plurality of pairs consisting of a feature set and respectively associated preset thresholds of at least one energy band of the direct conversion X-ray detector, the feature set comprising at least one measured feature.
10. The method according to any one of the preceding claims, wherein in a further step for image data recording by means of a medical imaging device the determined threshold value is set (S4) at the direct conversion X-ray detector.
11. A calculation unit (12) for determining a threshold value of at least one energy band of a direct conversion X-ray detector (9) of a medical imaging device (1), having means (13, 16, 22, 23) for performing the method according to any one of claims 1 to 10.
12. the calculation unit of claim 11, being connected with a medical imaging device comprising the direct-conversion X-ray detector, the calculation unit comprising a control unit (22) configured for setting the determined threshold at the direct-conversion X-ray detector.
13. An X-ray detector (9) in the form of a direct-conversion X-ray detector comprising a calculation unit (12) according to claim 11.
14. A computer program having a program code for performing the method according to any one of claims 1 to 10 when the computer program runs on a computer.
15. a computer-readable data carrier (13) having a program code of a computer program for performing the method according to any one of claims 1 to 10 when the computer program runs on a computer.
16. A medical imaging device (1) comprising a direct-conversion X-ray detector (9) and a calculation unit (12) according to claim 11 or 12.
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE102018208955.8 | 2018-06-06 | ||
DE102018208955.8A DE102018208955A1 (en) | 2018-06-06 | 2018-06-06 | Determining a threshold value for at least one energy band of a direct-converting detector |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110568471A true CN110568471A (en) | 2019-12-13 |
CN110568471B CN110568471B (en) | 2023-12-01 |
Family
ID=68651551
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910491851.3A Active CN110568471B (en) | 2018-06-06 | 2019-06-06 | Method for determining a threshold value of an energy band, computing unit and medical imaging device |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN110568471B (en) |
DE (1) | DE102018208955A1 (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113177907A (en) * | 2020-01-27 | 2021-07-27 | 西门子医疗有限公司 | Controlling a medical X-ray apparatus |
Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1911174A (en) * | 2005-08-03 | 2007-02-14 | 西门子公司 | Operating method for an image-generating medical engineering assembly and articles associated herewith |
CN102297873A (en) * | 2011-05-03 | 2011-12-28 | 杭州一二八医院 | Method for identifying cancer cell images by soft X-ray microscopic imaging |
CN103654828A (en) * | 2012-09-13 | 2014-03-26 | 西门子公司 | X-ray system and method to generate image data |
US20140270073A1 (en) * | 2013-03-12 | 2014-09-18 | Martin Spahn | Method and System for Acquiring an X-Ray Image |
DE102014102080A1 (en) * | 2014-02-19 | 2015-08-20 | Carl Zeiss Ag | Method for image acquisition and image acquisition system |
CN105122085A (en) * | 2013-10-09 | 2015-12-02 | 皇家飞利浦有限公司 | Method and device for generating an energy-resolved x-ray image with adapted energy threshold |
CN105916283A (en) * | 2015-02-19 | 2016-08-31 | 西门子股份公司 | Automatically determining an adjustment setting for a signal analysis parameter of an x-ray detector |
US20170000431A1 (en) * | 2015-06-30 | 2017-01-05 | Martin Spahn | Method for receiving energy -selective image data, x-ray detector and x-ray system |
DE102016215105A1 (en) * | 2016-08-12 | 2017-09-07 | Siemens Healthcare Gmbh | Method and data processing unit for determining an acquisition data processing algorithm |
US20170316562A1 (en) * | 2016-04-28 | 2017-11-02 | Siemens Healthcare Gmbh | Determining at least one protocol parameter for a contrast agent-assisted imaging method |
CN107405123A (en) * | 2014-11-06 | 2017-11-28 | 西门子保健有限责任公司 | Retrieved using the scan data of depth transducer data |
CN107492090A (en) * | 2016-06-09 | 2017-12-19 | 西门子保健有限责任公司 | Analyzed according to generated data using the tumor phenotypes based on image of machine learning |
CN107865657A (en) * | 2016-09-22 | 2018-04-03 | 西门子保健有限责任公司 | The method for automatically selecting the process of measurement that magnetic resonance examination is carried out with magnetic resonance device |
US20180122082A1 (en) * | 2016-11-02 | 2018-05-03 | General Electric Company | Automated segmentation using deep learned priors |
-
2018
- 2018-06-06 DE DE102018208955.8A patent/DE102018208955A1/en active Pending
-
2019
- 2019-06-06 CN CN201910491851.3A patent/CN110568471B/en active Active
Patent Citations (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1911174A (en) * | 2005-08-03 | 2007-02-14 | 西门子公司 | Operating method for an image-generating medical engineering assembly and articles associated herewith |
CN102297873A (en) * | 2011-05-03 | 2011-12-28 | 杭州一二八医院 | Method for identifying cancer cell images by soft X-ray microscopic imaging |
CN103654828A (en) * | 2012-09-13 | 2014-03-26 | 西门子公司 | X-ray system and method to generate image data |
US20140270073A1 (en) * | 2013-03-12 | 2014-09-18 | Martin Spahn | Method and System for Acquiring an X-Ray Image |
CN105122085A (en) * | 2013-10-09 | 2015-12-02 | 皇家飞利浦有限公司 | Method and device for generating an energy-resolved x-ray image with adapted energy threshold |
DE102014102080A1 (en) * | 2014-02-19 | 2015-08-20 | Carl Zeiss Ag | Method for image acquisition and image acquisition system |
CN107405123A (en) * | 2014-11-06 | 2017-11-28 | 西门子保健有限责任公司 | Retrieved using the scan data of depth transducer data |
CN105916283A (en) * | 2015-02-19 | 2016-08-31 | 西门子股份公司 | Automatically determining an adjustment setting for a signal analysis parameter of an x-ray detector |
US20170000431A1 (en) * | 2015-06-30 | 2017-01-05 | Martin Spahn | Method for receiving energy -selective image data, x-ray detector and x-ray system |
US20170316562A1 (en) * | 2016-04-28 | 2017-11-02 | Siemens Healthcare Gmbh | Determining at least one protocol parameter for a contrast agent-assisted imaging method |
CN107492090A (en) * | 2016-06-09 | 2017-12-19 | 西门子保健有限责任公司 | Analyzed according to generated data using the tumor phenotypes based on image of machine learning |
DE102016215105A1 (en) * | 2016-08-12 | 2017-09-07 | Siemens Healthcare Gmbh | Method and data processing unit for determining an acquisition data processing algorithm |
CN107865657A (en) * | 2016-09-22 | 2018-04-03 | 西门子保健有限责任公司 | The method for automatically selecting the process of measurement that magnetic resonance examination is carried out with magnetic resonance device |
US20180122082A1 (en) * | 2016-11-02 | 2018-05-03 | General Electric Company | Automated segmentation using deep learned priors |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113177907A (en) * | 2020-01-27 | 2021-07-27 | 西门子医疗有限公司 | Controlling a medical X-ray apparatus |
Also Published As
Publication number | Publication date |
---|---|
CN110568471B (en) | 2023-12-01 |
DE102018208955A1 (en) | 2019-12-12 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110114834B (en) | Deep learning medical system and method for medical procedures | |
CN110121749B (en) | Deep learning medical system and method for image acquisition | |
US7796795B2 (en) | System and method for computer aided detection and diagnosis from multiple energy images | |
US9480443B2 (en) | Method for selecting a radiation form filter and X-ray imaging system | |
US10755407B2 (en) | Systems and methods for capturing deep learning training data from imaging systems | |
JP2007524438A (en) | Compensation method in radiological image processing technology | |
US20220313176A1 (en) | Artificial Intelligence Training with Multiple Pulsed X-ray Source-in-motion Tomosynthesis Imaging System | |
CN105962959A (en) | Creating a resultant image for a specifiable, virtual X-ray quanta energy distribution | |
KR20180044421A (en) | Systems and methods for medical imaging of patients with medical implants for use in orthodontic surgery planning | |
CN101147158B (en) | Methods and systems for analyzing bone conditions using mammography device | |
KR20220137215A (en) | Apparatus and method for conversion of medical image based on artificial intelligence | |
WO2019200349A1 (en) | Systems and methods for training a deep learning model for an imaging system | |
Seo et al. | Deep focus approach for accurate bone age estimation from lateral cephalogram | |
US20220101526A1 (en) | Method for generating a trained machine learning algorithm | |
CN110568471B (en) | Method for determining a threshold value of an energy band, computing unit and medical imaging device | |
WO2019200351A1 (en) | Systems and methods for an imaging system express mode | |
KR20200086406A (en) | System for detecting bone tumour | |
WO2023100926A1 (en) | Medical information processing method and medical information processing device | |
WO2019200346A1 (en) | Systems and methods for synchronization of imaging systems and an edge computing system | |
CN115910318A (en) | Computer-implemented method and determination device for determining an abnormal structure | |
Holst et al. | An independent evaluation of a new method for automated interpretation of lung scintigrams using artificial neural networks | |
Ohlsson et al. | Automated decision support for bone scintigraphy | |
Zhang et al. | Automated detection of z-axis coverage with abdomen-pelvis computed tomography examinations | |
JP6956514B2 (en) | X-ray CT device and medical information management device | |
Nelson et al. | Can convolutional neural networks identify external carotid artery calcifications? |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |