Detection of bad detectors in idle state

Abstract

A fault checker system for an X-ray detector, comprising: an input interface (IN) for receiving a reading taken by a target detector pixel not exposed to X-radiation. A converter (CV) is configured to convert the reading into a metric. A thresholder (CP) is configured to compare the metric to at least one threshold and provide an indication as to whether the detector pixel is faulty based on the comparison.

Description

Technical Field

[0001] The present invention relates to a fault checker system for an X-ray detector, a method for fault checking an X-ray detector, an imaging device, a computer program unit, and a computer-readable medium. Background Technology

[0002] In some X-ray imaging systems (“imagers”) such as computed tomography scanners (“CT”), the X-ray detection system typically includes thousands of detector channels, also known as pixels.

[0003] It is possible for a channel to fail during the lifespan of a CT scanner. A failed channel is sometimes referred to as a bad detector (“BD”) (channel or pixel).

[0004] If data acquired at BD is used in image reconstruction, artifacts may occur.

[0005] The article "Image correction and Reconstruction for BreastBiopsy" by Tortajada et al. (July 20, 2008, Digital Mammography, LNCS, Springer Berlin Heidelberg, pp. 545-552) discloses a method for enhancing image quality, which includes the detection and correction of dead pixel effects.

[0006] Patent applications CN110118990A and US2013 / 126742A1 disclose a radiation detector that includes dead pixel correction. Summary of the Invention

[0007] Therefore, improvements to X-ray imaging are needed.

[0008] The object of the invention is achieved through the subject matter of the independent claims, wherein further embodiments are incorporated in the dependent claims. It should be noted that the aspects of the invention described below are equally applicable to methods for fault diagnosis of X-ray detectors, imaging apparatuses, computer program units, and computer-readable media.

[0009] According to a first aspect of the present invention, a fault checker system for an X-ray detector is provided, comprising:

[0010] An input interface for receiving readings acquired by the pixels of a target detector that are not exposed to X-ray radiation;

[0011] A converter configured to perform a conversion operation to convert the reading into a measurement; and

[0012] A thresholder is configured to compare the metric with at least one threshold and, based on the comparison, provide an indication of whether the detector (in particular the target detector pixel) is faulty.

[0013] Readings collected from the target pixel during the detector's idle state represent noise measurements. The nature of the noise, such as its fluctuations or other noise patterns, can be used here to identify faulty pixels. The fault may be located in the detector's circuitry, particularly in the channels of the target detector.

[0014] In one embodiment, the conversion operation performed by the converter includes a normalization operation applied to the readings. This allows for thresholding and thus makes fault finding more robust.

[0015] In one embodiment, the normalization operation correlates readings from a group of one or more other pixels with the reading acquired by the target pixel. This improves robustness and, in particular, allows for compensation of drift effects.

[0016] In one embodiment, a group of pixels is adjacent to the target pixel. However, readings from pixels elsewhere on the detector array can be used. In particular, all remaining pixels or only a portion thereof can be used. In another embodiment, the target pixel is included within a detector patch, and adjacent pixels are confined to the detector patch. This allows for better robustness in fault finding.

[0017] In one embodiment, the metric is configured to quantify fluctuations in the collected readings.

[0018] In an embodiment, the metric includes one or more of the following: i) an estimate of the standard deviation over time, ii) the sum of the absolute differences over time.

[0019] In one embodiment, the normalization operation includes forming a spatial median of readings in the group for one or more pixels.

[0020] The fault checker can run once or more during the interval between scans while the scanner XI is idle. The fault checker can also run quasi-continuously (at a relatively high repetition frequency) while the imager is idle. The fault checker is capable of detecting faulty detector pixels.

[0021] As used herein, "idle state" refers to a state in which the X-ray source is exposed without the detector. Specifically, this can be achieved by turning off the X-ray source from the power supply or otherwise disabling the X-ray source to prevent the detector from being exposed to X-ray radiation. Although less preferred herein, the detector can be removed from the X-ray beam. However, turning off the X-ray source is preferred herein, and especially keeps it off for a preset time between imaging duty cycles. Although the X-ray source remains off in the idle state, the detector itself remains "on," i.e., the detector remains powered while the fault checker performs a fault check. The detector remains powered on in the proposed idle state for fault checking so that readings can be generated and provided at the detector's output interface.

[0022] If the fault checker identifies a new BD pixel, that BD pixel can be processed before the next clinical scan. The fault checker does not require the application of external signals, such as X-rays.

[0023] In embodiments, the fault checker enables the detection of faults (fault modes), such as interruptions between a transducer stage (e.g., a photodiode) and digitization circuitry in a detector. Some detector circuitry includes operational amplifiers for a given channel. In embodiments, the effect of the impedance of the transducer stage (e.g., a photodetector array) on the standard deviation of the bias of the associated operational amplifier is used to detect conductor line interruptions. The effect may include measurable noise reduction. However, a damaged conductor line is not the only fault mode that the proposed fault checker can detect. Other detectable fault modes include short circuits that themselves manifest as increased noise.

[0024] In some embodiments, the fault checker can use the available data paths in the imager (especially in the detector) to collect readings. In some embodiments, additional circuitry may not be required.

[0025] With the proposed fault checker operating in idle mode, the imager can remain operational even if bad pixels are detected. A service call need not be triggered. Fault findings can be recorded for each pixel. Service calls can still be scheduled in the future while the imager remains operational until then. Downtime can be reduced or avoided in the process.

[0026] In one embodiment, the system may include a compensator that compensates for faulty channels during imaging based on the output of a fault checker that identifies bad pixels. The output of the fault checker may include a data structure, such as a table, that associates pixel addresses with flags to indicate whether a pixel is bad.

[0027] The fault checker system can operate fully automatically "in the background": without user intervention and without affecting clinical workflow.

[0028] One or more thresholds for a given (target) pixel can be dynamically adjusted, i.e., updated, based on readings from other pixels. In an embodiment, readings used for normalization can also be used for threshold adjustment. Threshold updates can be performed once in advance or repeatedly during a measurement cycle as new readings are collected. Threshold adjustments can be conditional on one or more specific conditions. For example, if the current measurement remains below (or, in an embodiment, above) the current threshold, no adjustment is performed in that cycle.

[0029] In another aspect, a method for fault diagnosis of an X-ray detector is provided, comprising the following steps:

[0030] Receive readings acquired by the pixels of the target detector that are not exposed to X-ray radiation;

[0031] Convert the readings into measurements; and

[0032] The metric is compared with at least one threshold, and based on the comparison, an indication is provided as to whether the detector is faulty (especially the target pixel).

[0033] In another aspect, an apparatus is provided, comprising:

[0034] X-ray imaging equipment; and

[0035] The system of any of the above embodiments.

[0036] In one embodiment, at least a portion of the system is integrated into the detector module of the X-ray imaging apparatus.

[0037] In this embodiment, the imaging device is a medical X-ray imaging device, particularly a computed tomography (CT) scanner.

[0038] In another aspect, a computer program unit is provided that, when run by at least one processing unit, is adapted to cause the processing unit to perform the method.

[0039] In another aspect, a computer-readable medium is provided on which the program unit is stored.

[0040] As used herein, “faulty” or “malfunctioning” specifically refers to any unintended configuration in the detector’s circuitry that could result in incorrect readings when the detector is used during imaging (on X-ray exposure). The detector circuitry may include, in particular, portions forming or affecting detector channels (also referred to herein as pixels). Such unintended configurations may include any one or more of the following: damaged conductor lines, short circuits, loose contacts, and others. Damaged conductor lines may include total or intermittent interruptions. Unintended configurations may also include faulty electronic components in the circuitry, such as capacitors, photodiodes, resistors, or others. Attached Figure Description

[0041] Exemplary embodiments of the invention will now be described with reference to the following figures, which are not to scale, wherein:

[0042] Figure 1 A perspective view of an X-ray imaging apparatus is shown;

[0043] Figure 2A A schematic block diagram of an X-ray detector module is shown.

[0044] Figure 2B An enlarged plan view of the two-dimensional X-ray detector module is shown;

[0045] Figure 3 A simplified circuit diagram of the pixel circuit in the X-ray detector module is shown.

[0046] Figure 4 A schematic block diagram of a fault diagnosis system for an X-ray detector module is shown.

[0047] Figure 5 A flowchart illustrating a method for troubleshooting an X-ray detector is shown; and

[0048] Figure 6 The illustration shows data collected by an X-ray detector and based on... Figure 5 A plot of exemplary measurement data processed by the method described in the document. Detailed Implementation

[0049] refer to Figure 1 The diagram shows a schematic perspective front view of a medical imaging system XI. The medical imaging system is preferably a rotating X-ray imaging system, such as a CT scanner. Other rotating imaging modalities, such as C-arm or U-arm X-ray imaging devices or mammography devices, are also contemplated herein. In short, any type of X-ray imager is contemplated herein, and the continued reference to the CT scanner-type imager XI is an exemplary embodiment and in no way limiting of this disclosure.

[0050] In particular, but not only, the CT imager XI may include a fixed gantry NG disposed in the examination chamber. The fixed gantry NG carries a rotating gantry RG rotatable about an examination area A about a rotation axis Z therethrough. The rotating gantry is annular and the examination area A is formed as an opening therein. An examination table TB extends at least partially into the examination area along the rotation axis Z (which may also be referred to herein as the imaging axis Z). The patient PAT or the object to be imaged resides on the examination table. The table TB, with the patient PAT on it, can advance along the imaging axis Z such that the region of interest is located in the examination area A. The examination table TB is optional.

[0051] The rotating gantry includes a detector module capable of detecting X-ray radiation. The rotating gantry RG may also include an X-ray source XS. The source XS may be arranged on the rotating gantry RG in a relative spatial relationship with the detector D and across the inspection area A.

[0052] During imaging, X-ray radiation is emitted from the X-ray source XS and interacts with the patient tissue, then exits from the distal end of the patient and strikes the detector D. The striking radiation is converted by the detector D into (projected) measurement data (sometimes called detector raw data). The measurement data collected at the detector D can represent intensity values. As envisioned herein, the imager XI includes a fault checker FC configured to detect faults in the detector module. The fault checker is configured to operate simultaneously with the imager XI in an idle state outside of imaging, i.e., simultaneously with the absence of X-ray radiation received by the detector. In particular, the X-ray source XS can be turned off simultaneously with the operation of the fault checker FC. The operation of the fault checker FC will be explained more fully below in Figures 2-6.

[0053] Continue to refer to Figure 1 During imaging, the gantry rotates, and the detector, and in this embodiment, the X-ray source, also rotates with it. Due to the rotation, measurement data can preferably be acquired from multiple different spatial orientations p relative to the patient. In some imaging protocols, the stage TB advances along the imaging axis Z to collect measurement data at different locations. The image planes (or “image domains”) from which image data can be reconstructed from the measurement data are schematically indicated by directions X and Y, each perpendicular to the imaging axis Z. Different parallel image planes exist, one for each location on the Z-axis. An external or onboard power supply (not shown) supplies power to the gantry RG (and / or components thereon) via a slip ring device. An operator console (not shown) allows the user (such a medical professional) to control the imaging operation. The user can use the operator console to issue imaging control signals, such as X-ray source settings, detector settings, or signals to control the speed of rotation, movement of the examination stage TB, etc.

[0054] Measurement data can be forwarded to an image processing system (IPS) via wireless or wired communication devices. The IPS can be configured as a computer system running imaging software such as image reconstruction algorithms that allow the conversion of (projected) measurement data from the projection domain into cross-sectional images in the image domain X, Y. Preferred multiple cross-sectional images can be obtained along the imaging axis Z, which can be assembled into a 3D image volume. Other tasks can be performed by the IPS. The IPS can reside on a single or multiple computers, such as in a “cloud” setup or other distributed architecture. Instead of or in addition to providing measurement data to the IPS, the measurement data can be forwarded for storage in a database (DB) (such as a PACS in an HIS) or other storage. The reconstructed images or measurement data can be visualized on a display device (MT) or can be processed in other ways.

[0055] Before moving on to the proposed fault tracker system FC in more detail, let’s first refer to Figure 2 to explain the operation of the X-ray detector module.

[0056] Figure 2A This is a schematic block diagram of the detector module, and Figure 2B Provides a plan view of the detector module as viewed from the source XS along the projection direction p (enlarged for illustrative purposes).

[0057] First refer to Figure 2A The detector module includes a data acquisition unit (DAQ). The basic function of the DAQ is to convert incoming X-ray radiation into an electrical signal representing the energy of X-ray photons. This electrical signal can be converted into digital signals by an analog-to-digital (A / D) circuit, and the digital signals can then be processed by an image processing system (IPS) into a time-domain image for, for example, visualization.

[0058] The data acquisition unit DAQ of detector D includes a transducer stage XC, in which X-ray photons are converted into electrical signals. The transducer stage XC can be configured as multiple spatially arranged pixel elements PX1, PX2. Figure 2B The planar view shows an example of a 2D pixel matrix layout with i columns and j rows. This view extends along... Figure 2B The projection direction p in the drawing plane. In some imaging modalities, such as in slotted mammography imagers, a 1D layout is also envisioned.

[0059] Each of pixels PX1 and PX2 is associated with and coupled to a corresponding pixel electronics PE1 and PE2. For illustration, only two pixels PX1 and PX2 are shown. For example, in some CT scanner detections, the number of such pixels can be as high as approximately 10. 5The magnitude is [not specified]. Pixel electronics PE1 and PE2 form a readout circuit that provides electrical signals in the analog domain to the shared electronics SE via readout lines. Each pixel has its own readout line. Electrical signals, voltages, or currents are provided via the readout lines as individual detector pixel readings, multiple times per unit time. The shared circuit SE or other downstream circuits can perform conditioning tasks, such as filtering, amplification, etc., and particularly include A / D circuitry, where analog signals, typically in voltage units, are converted into digital values, which can then be processed by the image processing system IPS into an image in the time domain. The output at the shared electronics stage SE forms the aforementioned measurement raw detector data π in the projection domain.

[0060] The proposed fault detector system is envisioned as being integrated into the circuitry of the data acquisition unit (DAQ), and in the embodiment, in the analog domain, as by Figure 2A The access points are shown in the diagram. Specifically, the analog signals provided by the corresponding pixel electronics PE1 and PE2 are processed by the fault checker FC to determine if a fault exists. Alternatively or additionally, the fault checker FC can be connected to the connection circuitry that couples the pixel elements PX1, PX2 in the transducer stage to the corresponding pixel electronics PE1, PE2. Generally, as used herein, "pixel" can refer to each individual assembly comprising a specific portion of the transducer stage (i.e., the pixel element) and its associated pixel electronics PE1, PE2 coupled to the pixel element. The connection from a given pixel to its readout line is also referred to herein as a "channel." There can be as many channels as there are pixels in the detector D.

[0061] In a preferred embodiment, and in addition to or instead of the above, the readings are collected by the fault checker not only in the analog domain but also in the digital domain, i.e., downstream (after) the A / D circuit.

[0062] Using these definitions, it can be said that a fault checker, as envisioned in this paper, is configured to identify faulty pixels, and the fault can occur anywhere in a given assembly or channel. Faulty pixels may be referred to herein as “bad detector pixels” or “bad pixel BD” or more simply as “pixel BD”.

[0063] In a preferred embodiment, the detector D is of the indirect conversion type, wherein the transducer stage XC includes a scintillator layer, typically made of crystal or any other suitable material such as ceramic or garnet. One side of the layer is arranged to face the incoming X-ray photons. Below the scintillator layer, i.e., on the other side, is an array of photodiodes PHDs, each photodiode preferably associated with a corresponding pixel location / element PX1, PX2 in the scintillator layer. In this embodiment, the scintillator is configured as a pixel scintillator element. The conversion is "indirect" because the incoming X-ray photons are first converted into visible light by the operation of the scintillator layer, and the photons in the visible spectrum are then converted into analog electrical signals by the photodiodes. Unstructured scintillators, such as single-crystal scintillators, are also contemplated, wherein there is no such one-to-one correspondence between the scintillator elements and the pixel electronics PE1, PE2. In this embodiment, pixels PX1, PX2 are defined by each of the individual pixel electronics PE1, PE2.

[0064] The fault diagnosis system and associated principles disclosed herein are not limited to indirect conversion type detectors. In particular, direct conversion type detectors are also conceived herein. In a direct conversion detector, the transducer stage XC is formed from a semiconductor material crystal such as a silicone layer, across which pixel electrodes are mounted as part of the pixel electronics. In other words, the electrodes for each pixel include an anode and a cathode. A voltage is applied across the semiconductor crystal by a power source (not shown). Incoming X-ray photons cause cloud charges to form. Cloud charges include holes and electrons. Electrons diffuse to the anode, while holes diffuse to the cathode, thus closing the circuit and emitting an electrical signal, which then travels along the corresponding readout line to the shared electronics SE.

[0065] In this embodiment, the detector module is assembled from individual sub-modules or pixel patches, rather than as... Figure 2B As shown, this forms a monolithic structure; however, it is also contemplated in other embodiments. Each block forms a group of pixels, and each pixel in a given block is typically served by a corresponding shared electronic circuit SE.

[0066] Typically, the shared electronic circuit SE can be implemented by an ASIC or other circuitry. Readout lines extending from the pixel electronic circuits PE1 and PE2 on a given chip are preferably coupled to the same shared electronic circuit, and readout is processed by this shared electronic circuit.

[0067] One failure mode leading to channel failure is a disconnection between the transducer stage XC, which converts X-ray energy into an electrical signal, and the input ports of the electronic circuitry in the pixel electronics PE1 and PE2. One effect proposed in this paper for identifying such bad channels is the increase in capacitance at the input of the charge integrating amplifier, which is measured by the amplifier output for noise. The transducer stage XC or its components add capacitance at the input of the amplifier that receives the electrical signal. Discontinuities reduce the capacitance at the amplifier input, and thus such discontinuities can be identified by a reduction in noise. Alternatively, another failure mode is a short circuit that can reduce the resistance at the amplifier input. Short circuits can be identified by the increase in noise they cause.

[0068] More details and references Figure 3 This illustrates a simplified wiring diagram of a portion of an integral-type indirect conversion detection circuit as contemplated in the embodiments herein. The diagram shows the individual pixel circuit PE1 and also includes a portion of its associated pixel element PX1. More specifically, in the embodiments, the circuit includes an integrating operational amplifier OA (“OpAmp”) and a photodiode PHD as part of the pixel element PX1, and another part as part of the scintillator layer. The measured value is forwarded from OpAmp to the readout line. In the embodiments, each pixel j and therefore each pixel electronic circuit PE... j It has its own readout line.

[0069] The circuit operates as follows. After a reset operation via switch SW, switch SW is opened to disconnect the reset capacitor RC from operational amplifier OA, initiating integration and generating a measurement reading. Current from photodiode PHD or bias BS flows to the input port of operational amplifier OA. This current causes the integrating capacitor C... int (It carries a negative charge) discharges. Once the integrating capacitor C... int Upon discharge, the operational amplifier OA triggers a new recharge cycle. Therefore, the number of recharge cycles per unit time is measured by the current at the input of OA, thus generating (noise) measurement results. Specifically, in the idle state (no X-ray exposure), the photodiode PHD can be described as a capacitor at the input of the operational amplifier OA. The operational amplifier OA, through C... int The bias current BS is measured by integration. An ideal operational amplifier has infinite input impedance. The capacitance of the photodiode PHD at the input reduces the input impedance and increases the noise in the bias measurement results. If the photodiode PHD is disconnected in some channels, the noise in the bias measurement results will be reduced for that channel. In particular, the scheme implemented by the fault checker FC allows the detection of interruptions in the idle state. The location of the interruption (such as a broken or otherwise damaged connection) is in... Figure 3The symbol shown as a cross "x" indicates other possible locations for fault interruptions between the output of the photodiode PHD and the input of the operational amplifier OA.

[0070] Other failure modes include short circuits. This occurs when an unexpected abnormal current path exists between the positive and negative inputs of operational amplifier OA (in...). Figure 3 A short circuit (indicated by "S") can occur, which can lead to a decrease in impedance between the inputs of the operational amplifier OA. The applicant has observed that the presence of a short circuit is associated with reduced noise. Short circuit S can also occur in other parts of the circuit.

[0071] The applicant has observed that both failure modes, intermittent and short-circuit, possess a unique “noise imprint” detectable in noise signal measurements when detector D is idle (no X-rays incident from source XR) to determine whether a given pixel is faulty. Another failure mode detectable in this paper is a fault within operational amplifier OA that can result in no output readings at all, and therefore zero noise.

[0072] Other failure modes can also be detected in this paper, and the examples above are not exhaustive of all failure modes. Furthermore, it will be understood that... Figure 3 The circuits described herein are solely for illustrating various failure modes and the noise modes they can induce. Variations of the circuits, and indeed other circuits, are also envisioned in this paper. Figure 3 The wiring diagrams herein should not be construed as limiting the principles described herein. However, some embodiments contemplated herein may still include the circuits shown or variations thereof, and in particular their circuit equivalents.

[0073] Now for reference Figure 4 It shows a schematic block diagram of a fault checker FC as envisioned in this paper.

[0074] The fault checker FC can be fully integrated into the imager XI, particularly into the detector module itself. The fault checker can be implemented on a single or multiple integrated circuits (“chips”). The fault checker can also be implemented on a microcontroller that includes processing unit storage devices and other components necessary for processing.

[0075] As briefly mentioned above, such as Figure 2AThe fault checker FC, indicated by a dotted line, can be connected to the analog domain circuitry of the detector, thus enabling it to process analog signals. In this case, the input port IN of the fault checker includes or is coupled to an analog-to-digital converter. Specifically, in embodiments, the fault checker FC is configured and arranged to collect its input as one or more analog signals in a segment of the DAQ circuit downstream (after) the photodiode PHD but upstream (before) the A / D conversion of the DAQ. However, and as also mentioned above, alternatively and preferably, the fault checker FC accepts its input after the digitization stage of the detector D, such as... Figure 2A The solid lines in the diagram illustrate this. Specifically, the fault checker FC can obtain its input from a conventional output interface, with the pixel readout lines terminated in the conventional output interface. This can be the same output that supplies the raw projection data π to the image processing system for reconstruction. In an embodiment, the fault checker FC is implemented via a suitably programmed processing unit PU. The processing unit PU may be dedicated to, or can be programmed to, perform other functions, such as controlling the transfer of data to the image processing system IPS, or others.

[0076] In short, the proposed fault checker FC is configured to operate in either the analog or digital domain, but operation in the digital domain is preferred. In either case, faults in analog and digital circuits can be detected using the proposed fault checker FC.

[0077] Some detectors D include a segment with logarithmic transformation. The logarithmic operation is applied to the output signal after A / D conversion to produce a "logarithmic" signal. The proposed fault checker FC is also configured to process logarithmic or non-logarithmic signals when checking for faults.

[0078] Broadly speaking, and continuing to refer to Figure 4 The fault checker receives input signals at input port IN, including pixel readings in the detector idle state. These signals can come from a single pixel, a group of pixels (such as patches), all pixels on the detector array, or other groups that are not necessarily topologically connected. The fault checker operates pixel-by-pixel and generates a corresponding indication of whether the corresponding pixel is faulty for each processed pixel signal. This indication for each pixel is output at output port OUT. The fault checker can operate in parallel on multiple (potentially all) pixels or groups of pixels, or it can process pixel readouts sequentially.

[0079] FC may include a digital high-pass filter (not shown) to be applied to the measured pixel signal to enhance noise. This is because it has been found that the spectral density of noise, such as that associated with the capacitance at the input of amplifier OA in the case of discontinuity, has a strong frequency dependence.

[0080] The operation of the fault checker FC involves a converter CV used to convert the pixel readout signal (analog or digital) into a metric, which will be explained in more detail below. The metric is then processed by a comparator CP, which compares the metric to one or more thresholds TH according to a test strategy or using other textual strategies. The comparison with one or more thresholds can be explicit or implicit. The thresholds can be fixed or can be dynamically changed over time for each pixel or some pixels and / or change from pixel signal to pixel signal for a given pixel. The comparator CP checks whether one or more thresholds have been broken according to the test strategy. This will be explained below. Figure 5 This section discusses some testing strategies.

[0081] In this embodiment, the comparator CP performs a fault test based on one or more thresholds. If the test fails, the corresponding pixel is considered bad, and this is indicated in an appropriate data structure, which is then output at port OUT. The per-pixel fault BD can be written to a log file or can be individually output as an appropriately coded flag, which is then processed / stored by the receiver. The log file can be stored in memory, on the fault checker FC board, or in external memory. The log file can be displayed on a display device MIT, or can be otherwise made available. The log file can be sent to the receiver, for example, for viewing by service personnel. The log file can be sent by the fault checker FC via an appropriate communication system.

[0082] In an embodiment, the fault checker FC also includes a fault compensator FCOMP. Even though faulty pixels have been detected, the fault compensator FCOMP allows the imager to continue operating. This is because only sporadic faults, determined by a certain critical number of faulty pixels, still allow the imaging device to operate safely and continuously. If a pixel is known to be bad, its reading can be discarded and replaced by interpolation from neighboring pixels identified as good. The fault compensator FCOMP can cooperate with an image processing system IPS, where measurement data is reconstructed into an image. Fault compensation can be carried over to the image domain or can be done in the projection domain. For example, measurement data collected in the projection domain during imaging operations can correct bad pixels by interpolation across neighboring measurements, and it is in this way that the projection data is "patched up," which is then fed into the image processing system for reconstruction. Alternatively, interpolation can be done in the image domain by modifying the contribution of bad pixels to a given voxel.

[0083] Therefore, although the output can be used to trigger a service call, this is not necessary in all cases, and the proposed FC ensures continued safe operation even if a faulty detector is detected. A service call is only issued when the total number of BD pixels exceeds a certain critical threshold or when too many BD pixels accumulate in a given area, making reliable robust interpolation impossible to guarantee. This can be done entirely automatically by the FC through an interface with an appropriate communication system. Messages including imager type, location, etc., can be sent to the service point to request a call.

[0084] Now for reference Figure 5 The flowchart illustrates a computer-implemented method for fault diagnosis of an X-ray detector. The steps in the flowchart are illustrated in more detail. Figure 4 The operation of the fault checker. However, it should be understood that the methods and steps described below constitute instruction in themselves and are not necessarily related to the operation of the fault checker. Figure 4 The architecture.

[0085] At step S510, the idle state of the X-ray imager XI is detected. This can be accomplished by an event handler that intercepts, for example, a switching signal issued by the user from an operator console associated with the imager XI. The switching signal may have already been issued by the user after the X-ray tube has been turned off by the user, specifically by operating the "Check Detector" button or other interface. Alternatively, once the event handler intercepts the switching signal for turning off the X-ray source, a fault check mode is automatically activated. Turning off the X-ray source may include interrupting power to the source. However, operating the collimator to block the X-ray beam can also trigger the detection of the idle state of the imager XI and may trigger a fault check.

[0086] More specifically, the idle state, as envisioned herein, particularly implies no X-ray radiation impacting the detector pixels (elements) PX1, PX2 to be fault-checked, and no scanning is expected within a preset time period. Preferably, the event triggering the fault-check mode is not only sent when the X-ray is off, since in some imaging protocols, the X-ray is turned off and on or collimated several times during the imaging process. Preferably, the fault-check trigger event then indicates not only that the X-ray is off, but also that the imager is not in an ongoing imaging process and / or such an imaging process is not expected to occur within the preset time period, such as in the next few minutes, the next half hour, the next hour, or any other suitable inactive period. This is to ensure that system performance is not degraded due to fault checking and parameter reset. A user interface, such as an overlay button or otherwise, may exist that allows the user to interrupt fault-check mode at any time and request a return to imaging mode. Alternatively, fault-check mode is automatically suspended when imaging mode is requested.

[0087] At step S520, the detector's detection mode (detection parameters) is appropriately set to now be processed in an idle state. Normally, detector pixel readings are not collected or analyzed in the idle state. However, in the proposed method, pixel readings are obtained outside the scan time. Setting the detection mode facilitates noise collection by setting the detection parameters to optimized values, i.e., values ​​that promote the collection of noise measurement results. In particular, and in embodiments, the detection parameters are set such that the noise added by the capacitance of the detection element (e.g., photodiode PHD) at the input of the detector channel can be identified based on the overall channel noise. In some (but not all) embodiments, the setting is specifically designed to account for the contribution of noise from the detectable PHD capacitor relative to the overall noise. Specifically, the setting is designed to reduce contributions from other noise sources. However, more specifically, the setting may include any one or more of the following: setting the gain of the detector channel to be processed, the integration time (for integration detection). Additionally or alternatively, other detection parameters are adjusted, such as the length of the time period during which a series of readings are collected, which is further shown in step S540.

[0088] At step S530, an analog signal, or preferably a digital signal, is measured for a given target pixel. The measurement result (“reading”) includes a noise measurement result, since there is no X-ray exposure due to the imager being idle. The proposed method will be explained below with an example of a specific given target pixel, and it is understood that the method is to be performed in the same manner on some or all pixels of the entire detector or a portion thereof (such as some or all pixels on a given detector patch or a portion thereof or on any other part of the detector array). The proposed method can be practiced in parallel for multiple pixels, or it can be practiced sequentially for some pixels. Even when practiced in parallel, this may include processing different groups of pixels sequentially.

[0089] Noise measurements can be obtained over a given period of time at a given frequency for a target pixel. The collection of pixel readings in the idle state in step S530 can be implemented by interfacing with an existing output port of a DAQ system for normal scanning in a busy operating state. Specifically, noise readings for any pixel can be obtained in the digital domain, as supplied by the readout circuit SE. Alternatively, the measurement signal can be collected in the analog domain, further upstream of the pixel electronics PE1 of the target pixel PX. The detector D, as envisioned herein, is configured such that readings are delivered even when the imager XI is in idle mode, ensuring that data can be collected in step S530 for fault-checking purposes. In other words, although the X-ray source can remain off during the execution of the proposed method, particularly during the measurement collection step S530, the detector itself remains powered on so that readings can be generated, as described above. Figure 3 As explained in the text. Step S530 can be implemented by one or more dedicated chips (such as microcontrollers or microprocessors) embedded in the detector module system.

[0090] At step S540, a metric is calculated based on the measurement results for the target pixel. The metric is either a suitably chosen statistical result or another quantity that captures the characteristics of noise fluctuations. Specifically, the measurement results for the target pixel include multiple readings taken for the target pixel over a given time series at step S530. This time series of noise measurement results is then used in step S540 to calculate a statistical metric, such as the sum of differences or standard deviations, their squares (variance), or other higher-order moments, such as third or fourth-order moments (skewness, kurtosis, respectively), or other values ​​such as autocorrelation, entropy, mixing distance, a combination of the foregoing, or any two or more, or any other suitable quantity (statistical or not). The metric is preferably suitable for quantifying fluctuations and / or noise, especially noise fluctuations.

[0091] In an embodiment of the metric, readings of the target pixel channels are collected over time and a statistical standard deviation is calculated. As previously stated... Figure 4 As mentioned, digital high-pass filters can be applied first to enhance the noise contribution in the measured pixel readings caused by the correlation between the frequency and spectral density of noise.

[0092] In optional step S550, the measurement is normalized. This may include considering readings collected over time at groups of one or more pixels other than the target pixel. The collection of measurement results for such pixels is the same as described above in step S530. A group may consist of pixels on the same patch, may consist of directly neighboring pixels, or may consist of any other group or groups in other (more distant) locations on the detector. A group of pixels may therefore be referred to as a reference pixel relative to the target pixel. Normalizing the measurement for a given target pixel based on the measurement results from the reference pixel (reference group) allows compensation for factors such as aging of the detector system. Using local (e.g., neighboring) pixels as a reference group allows for compensation for local effects. In particular, using some or all channels served by the same shared circuitry SE (e.g., ASIC) as a reference group allows for compensation for variations related to the ASIC's power supply or local temperature. Generally, using such channels that experience the same conditions as the target channel as a reference group makes thresholding more robust.

[0093] As mentioned above in step S520, another detection parameter that can be adjusted is the length of the period during which pixel readings are collected, and based on this, a measure such as the standard deviation is calculated. The collected pixel readings can be considered as a result of a random variable. Generally, for random variables, the accuracy of evaluating the standard deviation from the measurement results increases with the number of readings over time. However, a series of readings acquired over too long a period may be affected by low-frequency noise that can cause drift, which in turn may change the average reading of the channel and interfere with the ability of the measure to correctly capture noise. Normalizing the channel noise by using readings from other channels, as performed in step S550, can be used to compensate for such low-frequency drift. Normalization S550 can also be used to compensate for changes in environmental conditions and for aging of the detector system D.

[0094] In an embodiment, the array of detector pixels consists of groups (such as those pixels on a given chip) of pixels sharing the same electronic circuitry SE (such as a single ASIC for A / D conversion). For such groups of pixels served by the same shared electronics, similar or identical noise characteristics can be expected, and therefore similar values ​​for noise metrics, such as logarithmic readings or other mean absolute differences between them, can be expected. In this case, instead of using the entire pixel array of detector module D, normalization S550 can be performed on the channels served by the same ASIC. In such an embodiment, normalization can be performed per group (such as per chip).

[0095] At step S560, a testing strategy is applied. A testing strategy may include thresholding. At step S560, one or more thresholds are applied to the metric calculated in step S540, or its normalized version as obtained in step S550. Specifically, in an embodiment, thresholding may include checking whether the corresponding amount / metric for the target pixel PX1 is above or below a threshold. Two or more thresholds may be used to define a range and check whether the metric obtained in steps S540 and S550 is within or outside the range.

[0096] In step S570, a decision is made as to whether the target pixel PX1 is a BD (Browser Defect) based on the testing strategy. It will be understood that the testing strategy will depend on the semantics of the metric, i.e., on how the noise is measured. For example, in the embodiment mentioned above regarding the standard deviation as a metric, this can be compared to a low threshold, a high threshold, or both. A comparison with the low threshold can detect discontinuities at the input of the probe channel amplifier. Therefore, if the metric result is below the low threshold, the target pixel is a BD, and otherwise it is good. If both thresholds are used, it can be further checked that the metric is not too high, as this can also indicate a BD. Specifically, if the metric result is above the high threshold, this can identify a short circuit at the input of the probe channel amplifier or a fault in the amplifier or other electronic circuitry. Similar conclusions can be drawn using metrics other than the standard deviation.

[0097] In step S580, based on the decision made in step S570, the target pixel is identified as either a bad BD or a good (“OK”). This step may include constructing or writing an appropriate data structure, such as a table, in which pixel identifier entries are associated with corresponding flags (such as “BD” or “OK” or “1” / “0” or others) according to appropriate encoding. Alternatively, the list may include only BD pixel identifiers or only good pixel identifiers.

[0098] An optional step may exist, based on a data structure, to check for a critical number of bad pixels globally and / or within a specific neighborhood. If so, an alarm is issued and / or a service call is requested, and the imager leaves the service. If not, the imager is allowed to continue, and interpolation may be used to compensate for bad pixels if the imager exits the idle state and (re)enters regular imaging operation (“operational state”), which includes turning back on the X-ray source. It will be understood that the list (or other data structure, such as database entries) recording bad pixels is maintained during the busy state and is continued to be written to once the imager returns to the idle state again and the method is reapplied.

[0099] The method can be executed whenever the imager returns to an idle state. Alternatively, the method can be executed according to a schedule, such as when the imager is idle, or once per hour, per day, per week, etc., depending on explicit requirements. The method determines the state (BD or OK) for each pixel once or more when executed. Specifically, in one embodiment, the method is executed continuously at a given frequency, such as per second, while the imager XI is idle. Other periodic but slower schemes are also envisioned, such as evaluation per minute or per hour, while the imager XI is idle.

[0100] Typically, the proposed method can be executed for all pixels together, or the detector array can be processed portion by portion over time. Specifically, to conserve fault checker (FC) resources, each portion of the detector array is checked one after another in a loop until the entire array D has been fault-checked. Once all pixels in all detectors have been checked, the check loop restarts, where pixels in the portion of the detector processed first are (re)checked, and so on. If the fault checking method flow is interrupted, for example when an idle state cutoff is requested, the method flow will restart fault checking in the same portion of the detector array to the left of the interrupted state once the idle state is restored. This fault checking protocol ensures that, over time, the entire detector D will eventually be fault-checked.

[0101] It will be understood that the aforementioned processing will depend on the processor capabilities on which the fault checker is implemented. In embodiments, it may be possible, and is indeed preferred, to process the entire detector array in parallel. Preferably, the processor's processing mode (partial or together, serial or parallel) and processing power are respectively adjusted and selected such that a complete detector array fault check can be completed within a period of less than the average non-operation time between patient imaging at a given clinical site.

[0102] As envisioned in this document, steps S540 and S550 will now be explained in more detail.

[0103] The example implementation can be used for a rectangular detector array comprising 672 by 128 channels in an integrating type detector. The channels are set to an integration time of 300 μs. Multiple noise readings are collected, such as 4000 noise readings for a target channel without X-ray exposure.

[0104] By x i,j,n Refers to the nth reading of the channel at the i-th detector column and the j-th detector row, calculate the standard deviation of x along the reading.

[0105] The standard deviation of the noise of the i-th detector column and the j-th detector row can be represented as N. i,j N i,jThis represents the metric calculated at step S540 as described in the embodiment, and this can be compared with one or more thresholds to infer whether a pixel is a BD.

[0106] For better robustness, this document proposes, in a preferred embodiment, to normalize N at step S550 based on readings from a reference set of other channels. i,j In this embodiment, readings from all other channels are based on N forming rows and columns. i,j The quotient of the median is used to collect the data. Specifically, in the embodiment, the median of the noise along the column is calculated and used for normalization:

[0107] S i，j =N i，j / median 沿着指数i (N i，j (1)

[0108] Additionally, along S i,j The median of the rows is calculated and used for normalization to find the normalized noise:

[0109] N i,j =S i,j / median 沿着指数j (S i,j (2)

[0110] In the embodiments, the following variations of (1) and (2) are envisioned: it may not be necessary to collect the median over all rows or all columns, therefore i and j can be limited to cover, for example, only pixels for a given patch or any other group. In the embodiments, Figure 2B In the matrix layout, only four or eight intermediate nearest neighbors are considered for normalization, or slightly larger groups, such as squares whose centers include a given side length of the target pixel, etc. Other proximity geometries are also envisioned. In another alternative embodiment, instead of row and column midpoints as in (2), only column midpoints or only row midpoints are considered in the embodiment. Normalization based on quantities other than the midpoint can be used in (1), (2), such as average, weighted average, percentiles, etc.

[0111] In an alternative embodiment, instead of using the standard deviation N over time readings... i,j For example, by calculating the nearest neighbor difference (x i,j,n -x i,j,n-1 The standard deviation of the high-pass filter is used as a metric.

[0112] Normalized noise metrics such as (1) or (2) or other metrics can be compared with one or more thresholds in S560. The value of the (below) threshold can depend on the expected contribution of the capacitance of the probe element PHD to the overall noise. A typical value can be 0.85. Higher thresholds can also be applied. If, for example, the noise in the operating channel group is found to be within 0.85 to 1.2 times the median noise (or other average) of the pixels in the group, a high threshold of 1.5 times the median noise of the pixels in the group can be set.

[0113] The above embodiments (1) and (2) of calculating the noise measurement metric S540 and / or the normalization step S550 can be refined to consume less memory and / or are also applicable to the processing of logarithmic signals.

[0114] Storing the required number of readings per channel, such as 4000 readings per channel, for calculating noise according to, for example, (1) or (2), can impose a significant demand on system memory. It is desirable to use algorithms that allow the accumulation of data from multiple readings without significantly increasing the required memory. Another consideration is that in some systems, the digital output from the CT detector is the logarithm of the signal. It is advantageous to use algorithms capable of using either (non-logarithmic) or logarithmic signals as input.

[0115] As an alternative to (1), (2) is described below, which allows the accumulation of data from multiple readings to save memory and can accept similar performance, i.e., non-logarithmic or logarithmic signals.

[0116] As described in (1) and (2) above, the example implementation can be used for a rectangular detector array comprising 672 by 128 channels for an integration type detector. The channels can be set to an integration time of 300 μS or other values.

[0117] In step S530, Q readings are collected, for example, Q = 4001 readings for a target channel without X-rays. (By Lx) i,j,n Let Q-1 be the logarithms of the nth reading of the channel at the i-th detector column and the j-th detector row, and let Q-1 be the sum of the absolute values ​​of the differences between consecutive readings. This sum of absolute differences at the i-th detector column and the j-th detector row can then be written for this new metric as:

[0118]

[0119] In the normalization step S550, the median of the sums along the columns can be calculated and normalized as follows:

[0120] S i,j =s i，j / median 沿着指数i (s i,j (4)

[0121] Alternatively, or alternatively, calculations can be performed along S. i，j The median of the rows is normalized to find the normalized noise metric for this embodiment:

[0122] NN i,j =S i，j / median 沿着指数j (S i,j (5)

[0123] It will be recognized that for calculations (4) and (5), the logarithm of the signal can be used as input, but a (non-logarithmic) signal can be used instead. Furthermore, by summing the accumulated readings, only one memory cell is now required per channel, and a single reading counter common to all accumulated channels can be used, in contrast to the requirement of reserving n x m memory cells (where m is the number of readings for each of the n channels) as might be required for (1) and (2). Configurations (3)-(5) significantly reduce the required memory and computational complexity compared to calculating noise according to (1) or (2). Again, in (4) and (5), quantities other than the median can be used, such as other averages (possibly weighted) or other values.

[0124] In this embodiment, the summation operations in (1)-(5) are performed partly by the fault checker FC and partly by the common detector circuit SE or other detector D circuits. The fault checker FC can receive the partial sum from the detector D circuits as required by (1)-(5). This "outsourcing" partially reduces the amount of data to be transferred from the detector D to the fault checker FC and the load on the fault checker FC. Alternatively, all summation operations may be performed by the detector D electronics or by the fault checker FC.

[0125] Instead of the sum of differences in forming (5), a weighted sum of the differences can also be formed, wherein the corresponding weights w i Multiply by each or some of the addends in (4) and (5). In this way, the same reading can be given more weight than others, so that changes in the environment can be taken into account, for example, when obtaining the reading.

[0126] Throttling at S560 is as described above for (1) and (2): the normalized sum (5) is compared with one or more thresholds. The value of the lower threshold can depend on the expected contribution of the capacitance of the probe element to the overall noise. Again, a typical value can be 0.85. A higher threshold can also be applied. The variations of (1) and (2) described above also apply to (3)-(5).

[0127] Regarding one or more thresholds at step S570, these can be predefined and calculated from the design parameters of detector D (such as the characteristics of the photodiode PHD and / or amplifier OA or any other relevant electronic components, depending on the type of fault to be checked). Alternatively, the thresholds can be determined experimentally. As previously mentioned, values ​​less than 1 but greater than 0.5, such as 0.8, 0.85, 0.9, can be used for some detectors. Multiple thresholds can help identify multiple root causes or fault modes; for example, values ​​below 0.85 point to an intermittent fault at the photodiode PHD, while values ​​below 0.5 point to a fault in the operational amplifier OA. The thresholds can be fault-specific. It will be understood that the specific values ​​mentioned herein are exemplary, and other values ​​are also contemplated herein, but the mentioned values ​​may still be included in the embodiments.

[0128] Preferably, a dynamic threshold is used. Different threshold adjustment strategies can be used. In one embodiment, this is implemented by modifying a given predefined initial threshold based on the average of a metric for a group of reference pixels. In an embodiment, the metric can be the same as the metric used to measure noise in step S540, such as the sum of absolute continuous differences or the standard deviation over time.

[0129] More specifically, in an embodiment, the adjusted threshold may be set to the current threshold (which may initially be set to a predefined one) multiplied by the median or other average (possibly weighted) over a reference channel with the same metric, such as the sum of absolute differences or standard deviation. In other words, the threshold(s) may typically change over the measurement period. However, in other embodiments, the current threshold is only adjusted if the metric (such as the sum of absolute differences or standard deviation) for a given target channel is actually greater than the current threshold.

[0130] In a preferred embodiment, the data used for normalization calculations can be used to dynamically change the threshold using the same metric as described. The reference set used for calculating the normalization is the same as the reference set used for adjusting the threshold. However, this is not necessarily the case in all embodiments, where the sets may differ for the two purposes.

[0131] As an alternative to adjusting one or more thresholds in each measurement cycle S530, the threshold is fixed once and then maintained. Alternatively, the threshold is adjusted periodically, for example, when the imager XI is powered on or according to other schedules, once an hour, once a week, etc.

[0132] It should be understood that the above methods / algorithms can be practiced in many different embodiments, all of which are contemplated herein. Such embodiments may include different data acquisition schemes, such as collecting data from only a portion of the detector array D at a time, as previously mentioned. Another embodiment that can save memory is to store only some statistical properties of the data, such as the mean and variance of the overall or partial readings from each channel, instead of storing complete data from multiple readings. Statistical parameters from two or more groups of readings can be combined to improve the accuracy of noise estimation without significantly increasing memory usage.

[0133] Furthermore, it will be understood that the threshold decision strategy described above, i.e., whether results above or below a threshold are adopted to indicate that a target pixel is considered faulty, will depend on how the metric is calculated, the type and semantics of the metric, and possibly other design factors.

[0134] Although the processing of noise readings has been described above in the time domain, in some embodiments, the processing is performed in the frequency domain. The noise measurement is first transformed into a frequency signal in the spectrum using Fourier, Laplace, or wavelet transforms, and the frequency signal can then be processed similarly to what has been described above by calculating an appropriate metric capable of capturing fault-related time-domain fluctuations in the frequency domain or fault-specific noise imprints in the frequency domain. However, fault-checking analysis in the time domain as described herein is preferred.

[0135] Now for reference Figure 6 It shows a plot of a point cloud in a two-dimensional coordinate system. Figure 6 Drawing Figure 6 The efficiency of the proposed method has been demonstrated. Normalized noise was calculated as described above. Additionally, the signal Y with X-rays on was measured and normalized. Note that the signal measurement results recorded on the Y-axis with X-rays on are shown for illustrative purposes only and are generally not required when applying the method. Each channel in the detector array is composed of… Figure 6The dots in the diagram represent the channel's normalized noise and the Y-coordinate, respectively. The vertical dashed line corresponds to a normalized noise threshold of 0.85. The horizontal dashed line corresponds to an arbitrary threshold for "low signal" set for a normalized signal of 0.6. It can be seen that most channels with low normalized signals (below 0.6) are also below the normalized signal threshold (left side). Therefore, it is possible to identify most low-signal channels, i.e., BD pixels, from the noise measured as proposed in this paper with a high success rate. And this high success rate can still be achieved without X-ray exposure. It can be seen that only two low-signal channels are missed. These two low-signal channels can be caused by fault types other than interruptions or short circuits at the input of the channel amplifier. However, identifying such "abnormal" faults and setting thresholds and selecting metrics accordingly can lead to even further improvements in fault detection success rate.

[0136] Components of the fault checker FC can be implemented as one or more software modules running on a processing unit PU (such as a workstation associated with the imager XI), or associated with a server computer arranged in a distributed architecture and connectable to one or more imagers in a suitable communication network. Alternatively, some or all components can be arranged in hardware such as a suitably programmed microcontroller or microprocessor (such as an FPGA (Field Programmable Gate Array)), or as a hardwired IC chip, application-specific integrated circuit (ASIC) integrated into the detector module D or otherwise integrated into the imaging system XI. In yet another embodiment, the fault checker can be implemented using both software and hardware, or partly software and partly hardware. One or more features described herein can be configured or implemented as circuitry encoded in a computer-readable medium or configured or implemented using circuitry encoded in a computer-readable medium, and / or combinations thereof. Circuitry may include discrete and / or integrated circuits, system-on-a-chip (SoC) and combinations thereof, machines, computer systems, processors and memories, and computer programs.

[0137] In another exemplary embodiment of the invention, a computer program or a computer program unit is provided, characterized in that it is adapted to perform the method steps of the method according to one of the preceding embodiments on a suitable system.

[0138] Therefore, the computer program unit can be stored on a computer unit, which can also be part of an embodiment of the present invention. The computing unit can be adapted to perform the steps of the methods described above or to induce the execution of the steps of the methods described above. Furthermore, it can be adapted to operate components of the apparatus described above. The computing unit is adapted to automatically operate and / or execute user commands. The computer program can be loaded into the working memory of a data processor. The data processor can thus be configured to perform the methods of the present invention.

[0139] This exemplary embodiment of the invention covers both computer programs that use the invention from the outset and computer programs that convert existing programs into programs that use the invention through updates.

[0140] Furthermore, the computer program unit is capable of providing all the necessary steps of the process for implementing exemplary embodiments of the methods described above.

[0141] According to another exemplary embodiment of the invention, a computer-readable medium, such as a CD-ROM, is provided, wherein the computer-readable medium has computer program units stored on the computer-readable medium, the computer program units being described in the preceding portion.

[0142] Computer programs can be stored / distributed on suitable media (specifically, but not necessarily, non-transient media), such as optical storage media or solid-state media provided with or as part of other hardware, but computer programs can also be distributed in other forms, such as via the Internet or other wired or wireless telecommunications systems.

[0143] However, the computer program may also exist on a network such as the World Wide Web and can be downloaded from such a network to the working memory of a data processor. According to another exemplary embodiment of the invention, a medium is provided for making a computer program unit available for download, wherein the computer program unit is arranged to perform the method described in one of the previously described embodiments according to the invention.

[0144] It must be noted that embodiments of the present invention are described with reference to different subjects. Specifically, some embodiments are described with reference to claims of the method type, while others are described with reference to claims of the device type. However, those skilled in the art will understand from the above and below description that, unless otherwise indicated, any combination of features relating to different subjects, in addition to any combination of features belonging to one type of subject, is also considered to be disclosed in this application. However, all features can be combined to provide synergistic effects beyond the simple sum of the features.

[0145] Although the invention has been described and illustrated in detail in the accompanying drawings and the foregoing description, such description and illustration are to be considered illustrative or exemplary rather than restrictive. The invention is not limited to the disclosed embodiments. Other variations of the disclosed embodiments will be understood and implemented by those skilled in the art in practicing the claimed invention by studying the drawings, description, and dependent claims.

[0146] In the claims, the word "comprising" does not exclude other units or steps, and the words "a" or "an" do not exclude multiple. A single processor or other unit may perform the functions of several items recited in the claims. Although specific measures are recited in different dependent claims, this does not indicate that combinations of these measures cannot be advantageously used. Any reference numerals in the claims should not be construed as limiting the scope.

Claims

1. A fault checker system for an X-ray detector, comprising: an input interface (IN) for receiving a plurality of readings acquired over a time series of a given length by a target detector pixel that was not exposed to X-radiation; a converter (CV) configured to perform a conversion operation to convert the plurality of readings of the target detector pixel into a metric configured to quantify fluctuations in the acquired readings; and a thresholder (CP) configured to compare the metric to at least one threshold and provide an indication about whether the target detector pixel is faulty based on the comparison.

2. The system of claim 1, wherein, performing the conversion operation by the converter (CV) comprises the converter (CV) performing a normalization operation applied to the readings.

3. The system of claim 2, wherein, The normalization operation correlates readings from a group of one or more pixels with the readings acquired by the target detector pixel.

4. The system of claim 3, wherein, The group of pixels is adjacent to the target detector pixel.

5. The system of any one of claims 1 to 4, wherein, The metric comprises one or more of: i) an estimate of a standard deviation over time, ii) a sum of absolute differences.

6. The system of claim 3 or 4, wherein, The normalization operation comprises forming a spatial median for readings in the group of one or more pixels.

7. The system of any one of claims 1 to 4, wherein, The target detector pixel is comprised in a detector tile, and wherein the adjacent pixels are confined to the detector tile.

8. A method of fault checking an X-ray detector, comprising the steps of: receiving (S530) a plurality of readings acquired over a time series of a given length by a target detector pixel that was not exposed to X-radiation; converting (S540) the plurality of readings of the target detector pixel into a metric configured to quantify fluctuations in the acquired readings; and comparing (S570) the metric to at least one threshold and providing an indication about whether the target detector pixel is faulty based on the comparison.

9. An apparatus, comprising: an X-ray imaging apparatus; and the system of any one of claims 1-7.

10. The apparatus of claim 9, wherein, At least part of the system is integrated into a detector module of the X-ray imaging apparatus.

11. The apparatus of claim 9, wherein, The imaging apparatus is a medical X-ray imaging apparatus, in particular a computed tomography, CT, scanner.

12. A computer program element, which, when being executed by at least one processing unit (PU), is adapted to cause the processing unit (PU) to perform the method of claim 8.

13. A computer readable medium having stored thereon the program element of claim 12.

Images