A positron emission tomography method, system and device

By estimating and restoring the response line information of the abnormality detection unit in the PET device, the image artifacts and quantitative abnormalities caused by the damage to the detector module in the PET device are solved, and accurate image reconstruction and quality assurance in the case of damage are achieved.

CN114943782BActive Publication Date: 2024-11-26SHANGHAI UNITED IMAGING HEALTHCARE
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210566784.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-05-24
Publication Date
2024-11-26
Estimated Expiration
2042-05-24

AI Technical Summary

Technical Problem

The longer axial field of view in PET equipment causes the probability of detector module failure to increase, causing abnormal data, affecting the accuracy and quality of image reconstruction, especially the inability to effectively repair the flight time information when the detector module is damaged, resulting in image artifacts and Quantitative abnormality.

Method used

By acquiring the original data, the corresponding event count and time of flight information on the response line of the abnormal detection unit are estimated, and the TOF reconstruction algorithm chord diagram repair method is used to restore the information of the failure detector module and generate an accurate PET image.

Benefits of technology

The image can still be reconstructed correctly when the detector is damaged, ensuring image quantification and quality, reducing the probability of the reconstruction system downtime, improving robustness, and avoiding repeated scanning of the scanned object.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114943782B_ABST
    Figure CN114943782B_ABST
Patent Text Reader

Abstract

The embodiments of the present specification provide a positron emission tomography method, system and device. The method includes acquiring raw data of a scanned object. The raw data is obtained by a plurality of detection units of a PET detector, wherein the plurality of detection units include at least one abnormality detection unit. The method also includes estimating the count of the first coincident event on the response line corresponding to the at least one abnormality detection unit based on the raw data. The method also includes estimating the flight time information related to the response line corresponding to the at least one abnormality detection unit based on the raw data. The method also includes estimating the first flight time value of the first coincident event on the response line corresponding to the at least one abnormality detection unit based on the flight time information. The method further includes generating a PET image of the scanned object based on the raw data and the first flight time value of the first coincident event.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of positron emission tomography, and more specifically, to a method, system and device for reconstructing positron emission tomography images. Background Art

[0002] As a nuclear medicine imaging device, positron emission tomography (PET) equipment is widely used in the diagnosis, treatment and / or research of diseases (e.g., tumors). In recent years, the continuous increase in axial field of view has been the main development trend of PET equipment. On the one hand, a longer axial field of view brings about a huge improvement in detection sensitivity, which helps to further reduce the radiation dose received by the scanned object and improve the signal-to-noise ratio of the reconstructed image. However, on the other hand, a longer axial field of view corresponds to more PET detector modules (e.g., detection units), which greatly increases the probability of failure / abnormal modules (i.e., "bad channels") in the system. Bad channels can cause the collected PET raw data to include abnormal data (e.g., abnormal counts). If the raw data including abnormal counts are used to directly perform traditional PET image reconstruction, image artifacts and / or quantitative abnormalities will be caused, which will seriously affect clinical diagnosis.

[0003] Therefore, it is necessary to provide a method, system and device for repairing abnormal data in PET raw data, so as to more accurately reconstruct PET images and reduce reconstructed image artifacts, thereby improving the quality of reconstructed images. Summary of the invention

[0004] A first aspect of an embodiment of the present specification provides a positron emission tomography method. The method includes acquiring raw data of a scanned object. The raw data is obtained by detecting gamma photons by a plurality of detection units of a positron emission tomography detector, wherein the plurality of detection units include at least one abnormality detection unit. The method also includes estimating the count of a first coincident event on a response line corresponding to the at least one abnormality detection unit based on the raw data. The method also includes estimating the time-of-flight information associated with the response line corresponding to the at least one abnormality detection unit based on the raw data. The method also includes estimating the first time-of-flight value of a first coincident event on a response line corresponding to the at least one abnormality detection unit based on the time-of-flight information. The method further includes generating a positron emission tomography image of the scanned object based on the raw data and the first time-of-flight value of the first coincident event.

[0005] A second aspect of the embodiments of the present specification provides a positron emission tomography system. The system includes an acquisition module, a count estimation module, a flight time information estimation module, a flight time value estimation module and a reconstruction module. The acquisition module is used to acquire raw data of a scanned object. The raw data is obtained by detecting gamma photons by a plurality of detection units of a positron emission tomography detector, wherein the plurality of detection units include at least one abnormality detection unit. The count estimation module is used to estimate the count of a first coincident event on a response line corresponding to the at least one abnormality detection unit based on the raw data. The flight time information estimation module is used to estimate the flight time information related to the response line corresponding to the at least one abnormality detection unit based on the raw data. The flight time value estimation module is used to estimate the first flight time value of the first coincident event on the response line corresponding to the at least one abnormality detection unit based on the flight time information. The reconstruction module generates a positron emission tomography image of the scanned object based on the raw data and the first flight time value of the first coincident event.

[0006] A third aspect of the embodiments of this specification provides a positron emission tomography device, wherein the device comprises a processor, and the processor is used to execute the above-mentioned positron emission tomography method.

[0007] A fourth aspect of the embodiments of this specification provides a computer-readable storage medium, wherein the storage medium stores computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the above-mentioned positron emission tomography method. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0009] Figure 1 is an exemplary schematic diagram of a medical imaging system according to some embodiments of the present specification;

[0010] Figure 2 is an exemplary flow chart of a positron emission tomography method according to some embodiments of the present specification;

[0011] Figure 3 is an exemplary flow chart of recovering the count of coincidence events on the response line corresponding to the bad channel according to some embodiments of the present specification;

[0012] Figure 4is an exemplary flow chart of recovering the flight time information of the coincidence event on the response line corresponding to the bad channel according to some embodiments of the present specification;

[0013] Figure 5 is an exemplary flow chart of recovering the flight time value of a coincidence event on a response line corresponding to a bad channel according to some embodiments of the present specification;

[0014] Figure 6 is an exemplary flow chart of positron emission tomography according to some embodiments of the present specification;

[0015] Figure 7 are exemplary normal chord diagrams and bad channel chord diagrams according to some embodiments of the present specification;

[0016] Figure 8 are exemplary initial chord diagrams and repaired chord diagrams according to some embodiments of the present specification;

[0017] Fig. 9 is an exemplary module diagram of an imaging system according to some embodiments of the present specification. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] In the prior art, when the PET raw data includes abnormal data corresponding to a bad channel (eg, abnormal counts of response lines corresponding to the bad channel), the following methods are generally used to reconstruct the PET image.

[0020] The first method is to reconstruct the system modeling. In the system modeling, the influence of the line of response (LOR) related to the bad channel is removed, that is, it is assumed that the system itself does not have those detector modules corresponding to the bad channel. For example, the iterative formula of OP-OSEM can be shown as follows:

[0021]

[0022] Among them, a ij represents the system response matrix, i represents the number of the response line, and j represents the number of the reconstructed image pixel. In principle, all i∈I bad (i.e. the response line corresponding to the bad channel) ij Set to 0, because the natural count value g of the bad channeli is 0, so for the a in the second half of the above formula ij No special processing is required. Just replace the first half of ∑ i A i N i a ij That is, the sensitivity matrix can be modified. The specific method is to set the count on the corresponding LOR to 0 before the back-projection process of calculating the sensitivity matrix and then perform the back-projection.

[0023] The second method is chord repair. Without changing the reconstruction system modeling, the data that should be collected by the bad channel is estimated, and then the estimated data is added to the reconstruction data set for traditional reconstruction. For example, chord repair can be performed by difference. For another example, chord repair can be performed by deep learning (e.g., artificial intelligence (AI)).

[0024] Reconstruction system modeling can alleviate image artifacts to a certain extent by removing response lines related to bad channels, but it cannot repair quantitative deviations (for example, quantitative deviations in the metabolic process of tracers in the body). Although chord repair can remove image artifacts and restore image quantification to a certain extent, it often relies on interpolation or convolutional neural networks, so sufficient statistics are required on the chord. However, the amount of data collected in normal clinical practice can only guarantee the statistics on non-time of flight (non-TOF) chords; if the chord is rearranged into a time of flight (TOF) chord, the number of chord counts in each TOF dimension will drop by more than an order of magnitude, making interpolation or deep learning repair impossible. Therefore, chord repair can usually only be performed on non-TOF chords, and the flight time information (i.e., TOF information) of the recovery event cannot be obtained.

[0025] The embodiments of this specification mainly relate to a positron tomography method. Thanks to the improvement of the time characteristics of the PET system, the initial image of the scanned object can be directly obtained by directly "putting back" the time to the corresponding position of the response line according to the reconstruction algorithm of the flight time difference (TOF value). Then, the initial image is forward projected to obtain a guessed (or estimated) detector response chord diagram (TOF chord diagram) with TOF information, and the detector response chord diagram (non-TOF chord diagram) without TOF information obtained by chord diagram repair can be completely restored. All information of the events on the corresponding response line of the failed detector module (including the TOF value of the event and the event) can be fully restored, so as to reconstruct the image more accurately. Using the method disclosed in the embodiments of this specification, when the detector module of the PET detector is damaged, the PET image can still be correctly reconstructed, and the image quantification and quality are guaranteed to be not seriously affected (for example, the image quality and the accuracy of the image quantification are guaranteed to be within a certain range, avoiding the re-scanning of the scanned object to re-collect data, greatly reducing the probability of downtime of the PET system, and improving the robustness of the entire PET reconstruction system.

[0026] Figure 1 is an exemplary schematic diagram of an imaging system according to some embodiments of this specification. Figure 1 As shown, the imaging system 100 may include an imaging device 110 , a network 120 , one or more terminals 130 , a processor 140 , and a memory 150 .

[0027] In some embodiments, the imaging device 110 may include a positron emission tomography (PET) device, a single photon emission computed tomography (SPECT), etc. In some embodiments, the imaging device 110 may also include a computed tomography (CT) device, a magnetic resonance imaging (MRI) device, etc. For example, the imaging device 110 may include a PET-CT device. For another example, the imaging device 110 may include a PET-MRI device. For another example, the imaging device 110 may include a SPECT-CT device. For another example, the imaging device 110 may include a SPECT-MRI device. The imaging device 110 is used to scan a scan object (e.g., a patient) or a portion thereof to generate an image of the scan object or a portion thereof. The image may include a PET image, a PET-MR image, a PET-CT image, etc. For the purpose of description, the following description is made by taking the imaging device 110 as a PET device (referred to as PET device 110).

[0028] The PET device 110 includes a frame, a detector, a detection area, a scanning bed, etc. The axial field of view of the PET device refers to the axial direction (long axis direction of the scanning bed) of the PET device at one time. Figure 1 The range covered in the Z-axis direction (shown in the figure). Before the scanning process begins, a PET tracer is first administered to the scanned object. During the PET scan, the PET tracer emits positrons, which are antiparticles of electrons. Positrons have the same mass and opposite charge as electrons. Positrons and electrons (which naturally exist in the body of the scanned object) interact and annihilate (also called "annihilation events" or "coincidence events"). An electron-positron annihilation event produces two 511keV gamma photons (i.e., a pair of gamma photons), which travel in opposite directions relative to each other. The line connecting a pair of gamma photons is called a "Line of Response (LOR)". In some embodiments, the PET tracer may include a combination of one or more of the elements carbon (11C), nitrogen (13N), oxygen (15O), fluorine (18F), etc. For example, the PET tracer may be an organic compound labeled with one or more of the above elements.

[0029] Taking PET tracer as fluoroglucose (e.g., fluorine-labeled glucose) as an example, the aggregation degree of glucose in the scanned object is obtained by PET scanning. The area with more glucose aggregation means that the metabolism in the area is more vigorous, so the probability of the tumor being located in the area is higher. During the PET scanning process, after the scanned object absorbs fluoroglucose, an annihilation (i.e., an annihilation event or a coincidence event) will occur at the spatial position of the absorption, and then a pair of gamma photons will be emitted back to the two ends at the speed of light. The detector includes a plurality of detector modules arranged into one or more detector rings. An annihilation event or a coincidence event can be detected by two detector modules arranged relatively. The LOR of the annihilation event or the coincidence event passes through the two detector modules arranged relatively, and the pair of gamma photons corresponding to the annihilation event or the coincidence event can be detected by the two detector modules arranged relatively on the LOR. Each coincidence event detected by the PET detector corresponds to an LOR, which can be recorded by the two detector module numbers corresponding to the LOR, the radial distance r of the LOR, and the angle Φ between the LOR and the y-axis. After a PET scan, many coincidence events can be obtained. All coincidence events are arranged and coded according to the angle Φ and radial distance r of the coincidence response line to form a matrix diagram, namely the raw data sinogram (referred to as the chord diagram). The pixel value on the chord diagram is the count of the coincidence event corresponding to the position. The chord diagram without flight time information is called a non-TOF chord diagram. The chord diagram with flight time information is called a TOF chord diagram. In traditional PET technology, when a pair of γ photons are detected within the preset coincidence event window, it is considered that the annihilation time or coincidence event occurs. The specific position of the annihilation event or coincidence event in the LOR can only be determined through image reconstruction. Since there is no other information provided, the reconstruction algorithm will assume that the probability of annihilation events or coincidence events occurring at all positions on the LOR is the same when performing the initial reconstruction, that is, the contribution of each point on the LOR is the same, and then the annihilation events or coincidence events are counted according to the LOR to obtain the number of events on each LOR. This reconstruction process is called non-TOF reconstruction, and non-TOF reconstruction does not consider the flight time. It can be understood that for an LOR, every time a matching event corresponding to this LOR is detected, the count of matching events corresponding to each pixel on this LOR in the non-TOF chord diagram increases by 1. Therefore, the count of matching events corresponding to each pixel on the non-TOF chord diagram will be very large, which can ensure statistics.

[0030] As the time resolution of PET detectors increases, the time difference between a pair of γ photons arriving at two relatively set detector modules (also called the time of flight (TOF) value, i.e., TOF value) can be detected. The farther the annihilation position is from the center of the LOR, the greater the time difference between a pair of γ photons arriving at the two relatively set detector modules. According to the time of flight, the annihilation position of the annihilation event or coincidence event can be determined in theory. If the time resolution of the detector is very high, that is, the time of flight is sufficiently accurate (for example, accurate to 1 picosecond or 1 femtosecond), then the annihilation position of the annihilation event or coincidence event can be directly determined according to the time of flight for PET image reconstruction. But in fact, the time resolution of the detector is limited, so the time of flight may only be accurate to the order of hundreds of picoseconds, so the determined annihilation position of the annihilation event or coincidence event is not an exact point, but can only be limited to a certain range centered on the point. However, the reconstruction process can still be constrained based on this, the annihilation position is preliminarily determined within a certain range, and then a more reasonable weight probability allocation is made for the reconstruction information of the coincidence event. In this way, when reconstructing, it is assumed that only the voxels near the real annihilation position contribute to the projection value, which can reduce the number of voxels involved in the calculation and more reasonably calculate the contribution of each voxel to the projection value. This reconstruction process is called TOF reconstruction.

[0031] The network 120 may include any suitable network that can facilitate information and / or data exchange of the imaging system 100. In some embodiments, one or more components of the imaging system 100 (e.g., the imaging device 110, the terminal 130, the processor 140, the memory 150, etc.) may communicate information and / or data with one or more other components of the imaging system 100 via the network 120. For example, the processor 140 may obtain PET scan data from the imaging device 110 via the network 120. In some embodiments, the processor 140 may obtain user instructions from the terminal 130 via the network 120.

[0032] The terminal 130 may include a mobile device 131, a tablet computer 132, a laptop computer 133, etc., or any combination thereof. In some embodiments, the terminal 130 may be a part of the processor 140. In some embodiments, the terminal 130 may be used to input user instructions, display scan results, etc. In some embodiments, the terminal 130 may issue a prompt message to prompt the user. In some embodiments, the terminal 130 may be used to display image information (e.g., PET images, PET-CT images, PET-MR images, etc.).

[0033] The processor 140 may process data and / or information obtained from the imaging device 110, the terminal 130 and / or the memory 150. In some embodiments, the processor 140 may obtain raw data of the scanned object from the imaging device 110. In some embodiments, the processor 140 may estimate the count of the first coincident event on the response line corresponding to the abnormality detection unit based on the raw data. In some embodiments, the processor 140 may estimate the time of flight information (TOF information) related to the response line corresponding to the abnormality detection unit in the PET detector based on the raw data. In some embodiments, the processor 140 may estimate the time of flight value (TOF value) of the first coincident event on the response line corresponding to the abnormality detection unit based on the time of flight information. In some embodiments, the processor 140 may generate a PET image (i.e., a positron emission tomography image) of the scanned object based on the raw data and the time of flight value of the first coincident event.

[0034] The memory 150 may store data, instructions, and / or any other information. In some embodiments, the memory 150 may store data obtained from the terminal 130 and / or the processor 140. In some embodiments, the memory 150 may store data and / or instructions, and the processor 140 and / or the terminal 130 may execute or use the data and / or instructions to implement the exemplary methods described in this specification. In some embodiments, the memory 150 may be connected to the network 120 to communicate with one or more other components of the imaging system 100 (e.g., the processor 140, the terminal 130, etc.). One or more components of the imaging system 100 may access the data or instructions stored in the memory 150 through the network 120. In some embodiments, the memory 150 may be directly connected to or communicate with one or more other components of the imaging system 100 (e.g., the imaging device 110, the processor 140, the terminal 130, etc.). In some embodiments, the memory 150 may be part of the processor 140.

[0035] Figure 2 2 is an exemplary flow chart of a positron emission tomography imaging method 200 according to some embodiments of the present specification. The imaging method 200 may include steps 210, 220, 230, 240, and 250. In some embodiments, the imaging method 200 may be executed by the processor 140. In some embodiments, the imaging method 200 may be implemented by the imaging system 900 for PET scanning.

[0036] Step 210 , obtaining raw data of the scanned object. In some embodiments, step 210 may be implemented by the obtaining module 910 .

[0037] The raw data here refers to the PET scan data acquired by the imaging device (e.g., PET device) 110 performing a PET scan on the scanned object (e.g., a patient). For example, the PET device 110 can perform a continuous scan on the scanned object to obtain the raw data of the scanned object, that is, scan and obtain data of several time periods. For another example, the PET device 110 includes a PET detector. The PET detector includes a plurality of detector modules (also referred to as "detection units") to detect gamma photons (γ photons) to obtain the raw data of the scanned object. The plurality of detection units can be arranged to form one or more detection rings. The raw data may include counts and / or flight time information of coincident events on the response lines associated with the plurality of detection units. In some embodiments, the plurality of detection units include at least one abnormal detection unit. The abnormal detection unit refers to a damaged or failed detection unit (also referred to as a "bad channel"). The counts on the response lines (LORs) associated with the abnormal detection unit will be abnormal, for example, the counts cannot be collected or the collected counts are inaccurate. Thus, the abnormal detection unit will cause the raw data to include abnormal data obtained by detecting γ photons by at least one abnormal detection unit.

[0038] In some embodiments, the processor 140 may directly obtain the raw data of the scanned object from the PET device 110. In some embodiments, the processor 140 may obtain the raw data of the scanned object from the memory 150.

[0039] Step 220 , based on the raw data, estimate the count of the first coincidence event on the response line corresponding to at least one anomaly detection unit. In some embodiments, step 220 may be implemented by the count estimation module 920 .

[0040] Since the multiple detection units of the PET detector include at least one abnormal detection unit, the raw data includes the count of coincident events on the response line associated with the normal detection unit and the count of abnormalities including the coincident events (also called the first coincident events) on the response line associated with the abnormal detection unit. The abnormal count is zero or is not zero with a small probability.

[0041] In some embodiments, the processor 140 may obtain a first chord diagram of the scanned object based on the original data, and then repair the first chord diagram using difference and / or deep learning to obtain a second chord diagram to estimate the count of the first coincident event on the response line corresponding to at least one anomaly detection unit, thereby repairing the anomaly count of the first coincident event on the response line corresponding to at least one anomaly detection unit in the original data. For more descriptions on repairing the anomaly count, please refer to the description elsewhere in this specification (for example, Figure 3 and its related description).

[0042] Step 230 , estimating TOF information related to a response line corresponding to at least one abnormality detection unit based on the raw data. In some embodiments, step 230 may be implemented by a TOF information estimation module 930 .

[0043] Since at least one abnormal detection unit is a damaged or failed detection unit, the abnormal detection unit cannot detect the TOF information of the first coincident event on the response line related to the at least one abnormal detection unit (for example, the time when a pair of γ photons corresponding to the first coincident event are detected by the relatively set detection units) or the detected TOF information of the first coincident event is inaccurate, so that the original data includes the TOF information related to the response line corresponding to the normal detection unit, but does not include the TOF information related to the response line corresponding to the at least one abnormal detection unit, or includes abnormal TOF information related to the response line corresponding to the at least one abnormal detection unit.

[0044] In some embodiments, the processor 140 can obtain an initial image of the scanned object through TOF reconstruction based on the original data, and generate a third chord diagram by forward projecting the initial image. Based on the third chord diagram, TOF information related to the response line corresponding to at least one abnormal detection unit can be estimated. The TOF information may include the distribution of the flight time difference (i.e., TOF value) of the coincident events on the response lines corresponding to the multiple detection units (e.g., the proportion of coincident events with different flight time differences to the total number of coincident events). Both the first chord diagram and the second chord diagram are chord diagrams that do not include TOF information, i.e., non-TOF chord diagrams. The third chord diagram is a chord diagram that includes TOF information, i.e., a TOF chord diagram. The accuracy levels of the first chord diagram (or the second chord diagram) and the third chord diagram are different. For example, the first chord diagram (or the second chord diagram) corresponds to a first accuracy, and the third chord diagram corresponds to a second accuracy. The first accuracy is greater than the second accuracy, i.e., the third chord diagram is a relatively rough chord diagram, and the second chord diagram is a relatively clear and accurate chord diagram. In some embodiments, the pixels in the first chord diagram (or the second chord diagram) and the third chord diagram are accurate to the same level (e.g., centimeter level). For example, the first chord diagram (or the second chord diagram) and the third chord diagram are both grp4-level chord diagrams. For more descriptions on estimating TOF information, please refer to the descriptions elsewhere in this specification (for example, Figure 4 and its related description).

[0045] Step 240 , estimating a first TOF value of a first coincidence event on a response line corresponding to at least one abnormality detection unit according to the TOF information. In some embodiments, step 240 may be implemented by a TOF value estimation module 940 .

[0046] As described above, there is a difference in accuracy between the second chord diagram and the third chord diagram, so the third chord diagram cannot be accurately matched to the second chord diagram. For example, the second coincident event on the response line corresponding to at least one anomaly detection unit in the third chord diagram may not be able to correspond one-to-one with the first coincident event on the response line corresponding to at least one anomaly detection unit in the second chord diagram. For example, the number of second coincident events in the third chord diagram is 100, and the number of first coincident events in the second chord diagram is 98. In some embodiments, the processor 140 performs probability density sampling on each first coincident event based on the TOF information estimated in step 230 to determine the first TOF value corresponding to each first coincident event. For more descriptions on estimating TOF values, refer to the descriptions elsewhere in this specification (for example, Figure 5 and its related description).

[0047] Step 250 , generating a PET image of the scanned object according to the original data and the first TOF value of the first coincident event. In some embodiments, step 250 may be implemented by the reconstruction module 750 .

[0048] In some embodiments, the processor 140 may generate a fourth chordogram based on the first TOF value of the first coincident event and the raw data. For example, the processor 140 may obtain the TOF value of the coincident event on the response line corresponding to the normal detection unit based on the raw data. The processor 140 may obtain the estimated first TOF value of the first coincident event on the response line corresponding to at least one abnormal detection unit based on step 240. The processor 140 may obtain the count of coincident events on the response line corresponding to the normal detection unit and the estimated count of the first coincident event on the response line corresponding to at least one abnormal detection unit based on the second chordogram. The fourth chordogram is a repaired TOF chordogram, including the count and TOF value of coincident events on the response line corresponding to the normal detection unit and the estimated count and estimated first TOF value of the first coincident event on the response line corresponding to at least one abnormal detection unit. The processor 140 may rearrange the fourth chordogram based on the PET reconstruction requirements to generate a fifth chordogram. The pixels of the fifth chordogram and the pixels of the fourth chordogram are accurate to different levels. For example, the fourth chordogram is a chordogram accurate to the centimeter level, and the fifth chordogram is a chordogram accurate to the millimeter level. For another example, the fourth chord diagram is a grp4-level chord diagram, and the fifth chord diagram is a grp2-level chord diagram. In some embodiments, the processor 140 may rearrange the fourth chord diagram to generate list data.

[0049] Further, the processor 140 can perform image reconstruction based on the fifth chord diagram or the list data to generate a PET image of the scanned object. In some embodiments, different reconstruction algorithms can be used for image reconstruction, including analytical reconstruction algorithms or iterative reconstruction algorithms. Exemplary analytical reconstruction algorithms may include a filtered back projection (FBP) algorithm, a back projection filter (BFP) algorithm, a p-filter method, etc., or any combination thereof. Exemplary iterative reconstruction algorithms may include maximum likelihood expectation maximization (ML-EM), ordered subset expectation maximization (OSEM), row processing maximum likelihood (RAMLA), dynamic row processing maximum likelihood (DRAMA), etc., or any combination thereof.

[0050] It should be noted that the above description of the process of method 200 is for exemplary purposes only and does not limit the scope of this specification. For those skilled in the art, any changes or modifications can be made based on this specification. In some embodiments, the steps of method 200 are not sequential. In some embodiments, method 200 may include one or more additional steps or one or more steps of method 200 may be omitted. In some embodiments, at least two steps in method 200 may be combined into one step or one step in method 200 may be split into two steps. For example, steps 230 and 240 may be combined into one step.

[0051] Figure 3 300 is an exemplary flow chart of recovering the count of coincident events on the response line corresponding to the bad channel (i.e., at least one abnormality detection unit) according to some embodiments of the present specification. Process 300 may include steps 310, 320, and 330. In some embodiments, process 300 may be executed by processor 140. In some embodiments, each step in process 300 may be executed by count estimation module 920. In some embodiments, step 220 in method 200 may be implemented by process 300.

[0052] Step 310: Generate a first chord diagram based on the original data.

[0053] In some embodiments, the count of the first coincidence event on the response line corresponding to at least one anomaly detection unit in the first chord diagram is abnormal. Figure 7 As shown, the first chord diagram 720 is an exemplary bad channel chord diagram, and the first chord diagram 720 corresponds to the normal chord diagram 710. Compared with the normal chord diagram 710, the count anomaly of the first coincident event on the response line corresponding to at least one anomaly detection unit of the first chord diagram 720 is shown as one or more "black lines" of count anomaly with zero count (or non-zero count with a small probability).

[0054] In some embodiments, the first chord graph may be a compressed chord graph. The processor 140 may directly generate an initial chord graph based on the original data. The initial chord graph may be a chord graph accurate to the millimeter level, that is, the pixels on the initial chord graph are at the millimeter level. The processor 140 compresses the initial chord graph to generate the first chord graph. The first chord graph may be a chord graph accurate to the centimeter level, that is, the pixels on the first chord graph are at the centimeter level. For example, a 1 mm × 1 mm area on the initial chord graph corresponds to one pixel, and a 1 cm × 1 cm area on the first chord graph corresponds to one pixel. One pixel on the first chord graph corresponds to 10 × 10 (100) pixels on the initial chord graph. The reason why the original chord graph needs to be compressed is that the count of each pixel at the millimeter level is too small, and it is difficult to ensure statistics, so it is difficult to repair the first compliance event of at least one abnormality detection unit, and it will also lead to excessive calculation. For another example, the initial chord graph may be a grp2-level chord graph, and the first chord graph may be a grp4-level chord graph.

[0055] Step 320: Generate a second chord diagram based on the first chord diagram.

[0056] As described above, the first chord diagram includes the count of coincident events on the response line corresponding to the normal detection unit and the count of anomalies of the first coincident event on the response line corresponding to the abnormal detection unit, so it is necessary to repair the count of anomalies of the first coincident event on the response line corresponding to the abnormal detection unit. In some embodiments, the processor 140 can repair the first chord diagram by difference or deep learning to generate a second chord diagram. The second chord diagram includes the estimated count of the first coincident event on the response line corresponding to at least one abnormal detection unit. For example, based on the assumption that the chord diagram of instantaneous coincident events should be locally smooth, the first chord diagram should be locally smooth. When the area of ​​the abnormal detection unit accounts for a small proportion of the total area of ​​the multiple detection units of the PET detector, the missing part on the first chord diagram (i.e., the chord diagram pixel corresponding to at least one abnormal detection unit) is also correspondingly small. The processor 140 can obtain the second chord diagram by interpolating the count values ​​of the normal chord diagram pixels around the missing part on the first chord diagram. The second chord diagram includes the count values ​​on the chord diagram pixels corresponding to at least one abnormal detection unit (i.e., the count of the first coincident event on the response line corresponding to at least one abnormal detection unit). For another example, the processor 140 can obtain a chord repair model corresponding to the current PET detector (for example, directly obtain a pre-trained chord repair model or train to generate a chord repair model). The processor 140 inputs the first chord into the chord repair model, so that the chord repair model can output a second chord. The chord repair model is a machine learning model trained by a deep learning algorithm or an AI algorithm. As an example only, the processor 140 can obtain a normal chord generated by the current PET detector (for example, a normal chord generated when the PET detector does not have an abnormal detection unit). By shielding the area of ​​the corresponding part of the detection unit in the normal chord, a large number of abnormal chords with failed detection units (i.e., abnormal detection units) are obtained, and the abnormal chord and the corresponding normal chord are used as training sets. A chord repair model suitable for similar situations where the current PET detector fails is obtained by training using a machine learning model (for example, a convolutional neural network), thereby realizing the repair of data when the PET detector has an abnormal detection unit.

[0057] Step 330 : Estimate the count of first coincident events on the response line corresponding to at least one anomaly detection unit according to the second chord diagram.

[0058] In some embodiments, for each anomaly detection unit in at least one anomaly detection unit, the processor 140 uses the count on the response line corresponding to the anomaly detection unit restored in the second chord diagram as the estimated count on the response line corresponding to the anomaly detection unit.

[0059] It should be noted that the above description of process 300 is for exemplary purposes only and does not limit the scope of this specification. For those skilled in the art, any changes or modifications can be made based on this specification. In some embodiments, the steps of process 300 are not sequential. In some embodiments, process 300 may include one or more additional steps or one or more steps of process 300 may be omitted. In some embodiments, at least two steps in process 300 may be combined into one step or one step in process 300 may be split into two steps. For example, step 330 may be omitted.

[0060] Figure 4 4 is an exemplary flow chart for recovering the time of flight information (i.e., TOF information) of the response line corresponding to the bad channel (i.e., at least one abnormality detection unit) according to some embodiments of the present specification. Process 400 may include steps 410, 420, and 430. In some embodiments, process 400 may be executed by processor 140. In some embodiments, each step in process 400 may be executed by TOF information estimation module 930. In some embodiments, step 230 of method 200 may be implemented by process 400.

[0061] Step 410: Generate an initial image (also called histo-image) of the scanned object based on the original data.

[0062] In some embodiments, the processor 140 performs TOF reconstruction based on the raw data to generate an initial image of the scanned object. For example, the TOF information of the coincident event on the response line corresponding to the normal detector in the raw data can be used to ignore the time resolution of the detector to determine the theoretical annihilation position of the coincident event on the response line corresponding to the normal detector, thereby generating an initial image. The initial image is a blurred image, but it can reflect the general outline of the scanned object. Although the data of the response line corresponding to the abnormal detection unit may be missing, the response line of the normal detection unit will pass through the response line corresponding to the abnormal detection unit, so there will also be data on the response line corresponding to the abnormal detection unit on the initial image.

[0063] Step 420, performing a forward projection on the initial image to generate a third chord diagram.

[0064] The process of forward projection is the process of projecting the initial image from the image domain to the data domain. The initial image can reflect the spatial distribution of the metabolism of the tracer (e.g., fluoroglucose) in the scanned object, and then the chord diagram corresponding to the data collected by the PET detector that matches the event can be calculated by mathematical methods. For example, the processor 140 can project the initial image from the image domain to the data domain to generate the third chord diagram according to the following formula:

[0065]

[0066] Among them, y i represents the value of the i-th element (response line) in the third chord diagram, P ij represents the value of the system matrix for the jth image element (voxel) of the initial image and the i-th chord element, and represents the value of the jth element of the initial image. The system matrix reflects the contribution of the initial image to the response line, that is, the distribution of the probability of positron annihilation on the initial image.

[0067] Step 430 : estimating TOF information associated with a response line corresponding to at least one anomaly detection unit according to the third chord diagram.

[0068] Since the initial image has TOF information, the third chord diagram generated based on the forward projection of the initial image also has TOF information. For example, the third chord diagram can reflect the distribution of the flight time difference (also called "second TOF value") of the second conforming event on the response line corresponding to at least one anomaly detection unit (for example, the number of second conforming events corresponding to different second TOF values ​​or the proportion of the total number of second conforming events). In some embodiments, the processor 140 can obtain the distribution of the second TOF value of the second conforming event based on the third chord diagram. The processor 140 can use the distribution of the second TOF value of the second conforming event as the TOF information related to the response line corresponding to at least one anomaly detection unit. For more descriptions on estimating TOF information, please refer to the description elsewhere in this specification (for example, Figure 5 and related descriptions).

[0069] It should be noted that the above description of process 400 is for exemplary purposes only and does not limit the scope of this specification. For those skilled in the art, any changes or modifications can be made based on this specification. In some embodiments, the steps of process 400 are not sequential. In some embodiments, process 400 may include one or more additional steps or one or more steps of process 400 may be omitted. In some embodiments, at least two steps in process 400 may be combined into one step or one step in process 400 may be split into two steps. For example, step 430 may be omitted.

[0070] Figure 5500 is an exemplary flow chart for recovering the time of flight value (i.e., TOF value) of the response line corresponding to the bad channel (i.e., at least one abnormality detection unit) according to some embodiments of the present specification. Process 500 may include steps 510, 520, and 530. In some embodiments, process 500 may be executed by processor 140. In some embodiments, each step in process 500 may be executed by TOF value estimation module 940. In some embodiments, step 430 of process 400 and step 240 of method 200 may be implemented by process 500.

[0071] Step 510: Determine the count of second coincidence events corresponding to each second TOF value based on the third chord diagram.

[0072] In some embodiments, the processor 140 may determine a response line corresponding to at least one anomaly detection unit based on the third chord diagram. For a response line corresponding to each anomaly detection unit, the processor 140 may determine a count of second coincident events on the response line and a second TOF value corresponding to each second coincident event. The processor 140 may count the count of second coincident events corresponding to each second TOF value.

[0073] Step 520: Determine the ratio of the count of each second coincident event to the total count of second coincident events as the probability value of each second TOF value.

[0074] As an example only, assuming that the total count of second matching events is 100, the count of second matching events corresponding to the second TOF value of 0 picoseconds is 50, the count of second matching events corresponding to the second TOF value of 50 picoseconds is 30, and the count of second matching events corresponding to the second TOF value of 100 picoseconds is 20, then the processor 140 can determine that the probability value of the second TOF value being 0 picoseconds is 50%, the probability value of the second TOF value being 50 picoseconds is 30%, and the probability value of the second TOF value being 100 picoseconds is 20%.

[0075] Step 530: Determine TOF information related to a response line corresponding to at least one abnormality detection unit according to the probability value of each second TOF value.

[0076] In some embodiments, the processor 140 may determine the distribution of the second TOF value based on the probability value of the second TOF value on the response line corresponding to at least the abnormality detection unit. The processor 140 may determine the distribution of the second TOF value as TOF information related to the response line corresponding to at least one abnormality detection unit. That is, the TOF information includes the distribution of the second TOF value.

[0077] Step 540: for each first coincidence event, according to the TOF information, assign a second TOF value to the first coincidence event as the first TOF value of the first coincidence event.

[0078] In some embodiments, for each first conforming event, the processor 140 can perform probability density sampling for the first conforming event based on the TOF information, and use the second TOF value obtained by sampling as the first TOF value of the first conforming event. That is, the first TOF value of each first conforming event is determined based on the distribution of the second TOF value. For example, assuming that the distribution of the second TOF value is that the total count of the second conforming event is 100, the count of the second conforming event corresponding to the second TOF value of 0 picoseconds is 50, the count of the second conforming event corresponding to the second TOF value of 50 picoseconds is 30, and the count of the second conforming event corresponding to the second TOF value of 100 picoseconds is 20, then the processor 140 randomly extracts a second conforming event from the distribution. The processor 140 assigns the second TOF value of the randomly extracted second conforming event to the first conforming event as the first TOF value of the first conforming event. The probability value of the first TOF value of each first coincident event being assigned to 0 picoseconds is 50%, the probability value of the first TOF value being assigned to 50 picoseconds is 30%, and the probability value of the first TOF value being assigned to 100 picoseconds is 20%, that is, the distribution of the first TOF value of the first coincident event is consistent with the distribution of the second TOF value of the second coincident event.

[0079] It should be noted that the above description of process 500 is for exemplary purposes only and does not limit the scope of this specification. For those skilled in the art, any changes or modifications can be made based on this specification. In some embodiments, the steps of process 500 are not sequential. In some embodiments, process 500 may include one or more additional steps or one or more steps of process 500 may be omitted. In some embodiments, at least two steps in process 500 may be combined into one step or one step in process 500 may be split into two steps. For example, step 330 may be omitted.

[0080] Figure 6 600 is an exemplary flow chart of positron emission tomography (i.e., PET imaging) according to some embodiments of the present specification. Process 600 may include one or more steps of step 610, step 620, step 630, step 641, step 642, step 651, step 652, step 660, step 670, step 681, step 682, step 690, etc. In some embodiments, process 600 may be executed by processor 140. In some embodiments, process 600 may be implemented by imaging system 900 for PET scanning.

[0081] Step 610: Perform a PET scan on the scan object.

[0082] In some embodiments, the processor 140 may receive a scanning instruction, determine a scanning protocol or scanning parameters based on the scanning instruction, and control the imaging device 110 to perform a PET scan on the scanned object based on the scanning protocol or scanning parameters.

[0083] In step 620, it can be determined whether the PET detector has a failure detection unit (ie, an abnormal detection unit). When it is determined that the PET detector has an abnormal detection unit, step 630 is executed; when it is determined that the PET detector does not have an abnormal detection unit, step 690 is executed.

[0084] In some embodiments, the processor 140 may determine the current functional state of each of the plurality of detection units in the PET detector. For example, the processor 140 may determine the current functional state of the plurality of detection units based on one or more performance parameters. As an example only, one or more performance parameters of the detection unit are generated in real time during the scanning process of the PET detector. Exemplary performance parameters of the detection unit may include a count rate. The count rate of the detection unit may be determined by dividing the count of a single event counted by the detection unit in a certain time period by the time period. The count rate of the detection unit acquired in real time may be sent to the processor 140. The processor 140 may compare the acquired count rate with a count rate threshold. When the count rate of the detection unit is lower than the count rate threshold, the processor 140 may determine that the detection unit is an abnormal detection unit. When the count rate of the detection unit is equal to or exceeds the count rate threshold, the processor 140 may determine that the detection unit is a normal detection unit. In some embodiments, it may be determined whether the PET detector has an abnormal detection unit by some performance detection devices. The processor 140 may determine whether there is an abnormal detection unit and the specific location (e.g., number) of the abnormal detection unit based on the detection results of the performance detection device.

[0085] Step 630: Acquire the original data of the scanned object.

[0086] In some embodiments, the raw data includes an accurate count of the coincidence events on the response line corresponding to the normal detection unit, and an abnormal count of the first coincidence event on the response line corresponding to the abnormal detection unit. For more information about the raw data, see this specification. Figure 2 Step 210 of method 200 is shown and its associated description.

[0087] Step 641: Generate an initial image of the scanned object based on the original data.

[0088] In some embodiments, the processor 140 can reconstruct the raw data through TOF reconstruction to generate an initial image of the scanned object. The initial image is a relatively rough image that can show the outline of the scanned object. For more information about generating the initial image, see this specification. Figure 4 Step 410 of process 400 is shown and its related description.

[0089] Step 642, perform orthographic projection on the initial image to generate a third chord diagram.

[0090] In some embodiments, the third chord diagram is a chord diagram with TOF information. For example, the third chord diagram includes the TOF value distribution of the second coincident event on the response line corresponding to the anomaly detection unit. For more information about generating the third chord diagram, see this specification. Figure 4 Step 420 of process 400 is shown and its related description.

[0091] Step 651, generating a first chord diagram based on the original data.

[0092] In some embodiments, the count of the first coincident event on the response line corresponding to the anomaly detection unit in the first chord diagram is abnormal. In some embodiments, the processor 140 can generate an initial chord diagram based on the original data. The processor 140 can compress the initial chord diagram to generate a second chord diagram. For more information about generating the first chord diagram, please refer to the specification. Figure 3 Step 310 of process 300 is shown and its related description.

[0093] Step 652: Generate a second chord diagram based on the first chord diagram.

[0094] In some embodiments, the processor 140 repairs the first chord diagram by interpolation or deep learning to generate a second chord diagram. The second chord diagram includes an estimated count of the first coincident event on the response line corresponding to the anomaly detection unit. For more information about generating the second chord diagram, see the specification. Figure 3 Step 320 of process 300 is shown and its related description.

[0095] Step 660: Determine a first TOF value of each first coincident event in the second chord diagram by probability density sampling according to the TOF information in the third chord diagram.

[0096] In some embodiments, the processor 140 can determine the count and second TOF value of the second conforming event on the response line corresponding to the anomaly detection unit based on the third chord diagram. The processor 140 can count the count of the second conforming event corresponding to each second TOF value to determine the distribution of the second TOF values ​​of the second conforming event on the response line corresponding to the anomaly detection unit. The processor 140 can assign a second TOF value as the first TOF value of the first conforming event to each first conforming event in the second chord diagram through probability density sampling based on the distribution of the second TOF value. For more descriptions on determining the first TOF value of each first conforming event in the second chord diagram, please refer to this specification Figure 5 Flow 500 is shown and its associated description.

[0097] Step 670, generating a fourth chord diagram.

[0098] In some embodiments, the processor 140 may generate a fourth chord diagram based on the raw data and the first TOF value of the first coincident event. The fourth chord diagram is a TOF chord diagram, including an estimated count of the first coincident event on the response line corresponding to the anomaly detection unit and an estimated first TOF value. For more information about generating the fourth chord diagram, see the specification. Figure 2 The relevant description of step 250 of method 200 is shown.

[0099] Step 681, generating a fifth chord diagram based on the fourth chord diagram.

[0100] In some embodiments, the processor 140 may rearrange the fourth chord map to generate a fifth chord map according to the PET reconstruction requirements. The pixels of the fifth chord map are accurate to a first level (e.g., millimeter level). The pixels of the fourth chord map are accurate to a second level (e.g., centimeter level). For more information about the fifth chord map, see the specification. Figure 2 The relevant description of step 250 of method 200 is shown.

[0101] Step 682, generating list data based on the fourth chord diagram.

[0102] In some embodiments, the processor 140 may rearrange the fourth chord diagram to generate list data.

[0103] Step 690: Generate a PET image of the scanned object.

[0104] In some embodiments, when the PET detector does not have an abnormality detection unit, the processor 140 directly generates a PET image of the scanned object based on the original data of the scanned object. In some embodiments, when the PET detector has an abnormality detection unit, the processor generates the fifth chord diagram of step 681 or the list data of step 682 based on repairing the original data of the scanned object. The processor performs image reconstruction based on the fifth chord diagram or the list data to generate a PET image of the scanned object. For more information about image reconstruction, see this specification. Figure 2 A description of step 250 of method 200 is shown.

[0105] Figure 8 are exemplary initial chord diagrams and repair chord diagrams according to some embodiments of this specification. Figure 8 As shown, chord diagram 801-1, chord diagram 803-1, chord diagram 805-1 and chord diagram 807-1 are respectively initial chord diagrams obtained by rearranging the original data with abnormal counts collected by the PET detector. By respectively executing the chord diagram repair method described in this specification (for example, the relevant description of steps 210-250 and / or steps 610-690) on the above initial chord diagrams, the repaired chord diagram 801-2, chord diagram 803-2, chord diagram 805-2 and chord diagram 807-2 can be obtained. Among them, chord diagram 801-2 is a repaired chord diagram corresponding to the initial chord diagram 801-1, chord diagram 803-2 is a repaired chord diagram corresponding to the initial chord diagram 803-1, chord diagram 805-2 is a repaired chord diagram corresponding to the initial chord diagram 805-1, and chord diagram 807-2 is a repaired chord diagram corresponding to the initial chord diagram 807-1. By comparison, the artifacts in the repaired chord diagram are lower than the artifacts in the corresponding initial chord diagram.

[0106] Fig. 9 9 is an exemplary module diagram of an imaging system (e.g., a PET imaging system) according to some embodiments of the present specification. In some embodiments, the imaging system 900 may include an acquisition module 910, a count estimation module 920, a TOF information estimation module 930, a TOF value estimation module 940, and a reconstruction module 950. In some embodiments, each module in the imaging system 900 may be implemented by the processor 140.

[0107] The acquisition module 910 can be used to acquire information / data related to imaging. In some embodiments, the acquisition module 910 can be used to acquire raw data of the scanned object. For more information about raw data, see this specification. Figure 2 The relevant description of step 210 of the method 200 is shown. In some embodiments, the acquisition module 910 can acquire a scanning instruction to scan the scanning object.

[0108] The count estimation module 920 can be used to estimate the count of the first coincident event on the response line corresponding to at least one abnormal detection unit in the PET detector. In some embodiments, the count estimation module can estimate the count of the first coincident event on the response line corresponding to at least one abnormal detection unit based on the original data using an interpolation or deep learning scheme. For more information about count estimation, see this specification. Figure 2 A description of step 220 of method 200 is shown.

[0109] The TOF information estimation module 930 can be used to estimate the TOF information related to the response line corresponding to at least one abnormal detection unit. In some embodiments, the TOF information estimation module 930 can estimate the TOF information related to the response line corresponding to at least one abnormal unit through TOF reconstruction and forward projection based on the original data. For more descriptions on TOF information estimation, please refer to this specification. Figure 2 A description of step 230 of method 200 is shown.

[0110] The TOF value estimation module 940 can be used to estimate the first TOF value of the first conforming event on the response line corresponding to at least one abnormal detection unit. In some embodiments, the TOF value estimation module 940 can estimate the first TOF value of the first conforming event on the response line corresponding to at least one abnormal detection unit by probability density sampling based on TOF information. For more descriptions on TOF value estimation, see this specification. Figure 2 A description of step 240 of method 200 is shown.

[0111] The reconstruction module 950 can be used for image reconstruction. In some embodiments, the reconstruction module 950 can generate a PET image of the scanned object based on the original data and the first TOF value of the first coincident event. For more information about image reconstruction, see this specification. Figure 2 A description of step 250 of method 200 is shown.

[0112] In some embodiments, the functions of at least two or more modules of the imaging system 900 can be implemented by the same module. For example, the functions of the TOF information estimation module 930 and the TOF value estimation module 940 can be implemented by the same module. At least the imaging system 900 can also include one or more other modules. For example, the imaging system 900 may include a control module for controlling the imaging system 900 to scan the scanned object. For another example, the imaging system 900 may include a display module for presenting the PET image of the scanned object to the user.

[0113] The beneficial effects that can be achieved by the embodiments of the present specification may include but are not limited to: (1) the chord repair scheme and TOF reconstruction can be combined to count the raw data collected by the detector with a failed detection unit and restore the TOF information, so that image reconstruction can be performed more accurately; (2) when the detector has an abnormal detection unit, the image can still be correctly reconstructed, and the quantification and quality of the image are not seriously affected, reducing or avoiding rescanning and supplementary scanning of the scanned object, improving the user experience of the scanned object, reducing the probability of system downtime, and improving the robustness of the entire PET reconstruction system.

[0114] The embodiments of this specification are described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the process in the flowchart. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0117] Although the preferred embodiments of this specification have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of this specification.

[0118] Obviously, those skilled in the art can make various changes and modifications to the embodiments of this specification without departing from the spirit and scope of the embodiments of this specification. Thus, if these modifications and variations of the embodiments of this specification fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A positron emission tomography method, characterized in that: include: Acquiring raw data of the scanned object, the raw data being obtained by detecting gamma photons by a plurality of detection units of a positron emission tomography detector, wherein the plurality of detection units include at least one anomaly detection unit; estimating, based on the raw data, a count of first coincident events on a response line corresponding to the at least one anomaly detection unit; estimating, based on the raw data, time-of-flight information associated with a response line corresponding to the at least one anomaly detection unit; estimating, based on the flight time information, a first flight time value of a first conforming event on a response line corresponding to the at least one anomaly detection unit; A positron emission tomography image of the scanned object is generated according to the original data and the first flight time value of the first coincident event.

2. The method according to claim 1, characterized in that The estimating, based on the raw data, the count of the first coincident event on the response line corresponding to the at least one anomaly detection unit comprises: generating a first chord diagram according to the original data, wherein the count of coincident events on a response line corresponding to the at least one anomaly detection unit in the first chord diagram is abnormal; A second chord diagram is generated according to the first chord diagram, wherein the second chord diagram includes a count of the estimated first coincidence events on a response line corresponding to the at least one anomaly detection unit.

3. The method according to claim 1, characterized in that The estimating, based on the raw data, time-of-flight information associated with a response line corresponding to the at least one anomaly detection unit comprises: generating an initial image of the scanned object according to the original data; Performing a forward projection on the initial image to generate a third chord diagram, wherein the third chord diagram includes a second time-of-flight value of a second coincident event on a response line corresponding to the at least one anomaly detection unit; The time-of-flight information associated with a line of response corresponding to the at least one anomaly detection unit is estimated based on the third chord diagram.

4. The method according to claim 3, characterized in that The estimating, based on the third chord diagram, flight time information associated with a response line corresponding to the at least one anomaly detection unit comprises: determining a count of second coincident events corresponding to each of the second time-of-flight values ​​based on the third chord diagram; determining a ratio of the count of each second coincident event to the total count of second coincident events as a probability value of each second flight time value; The flight time information associated with the response line corresponding to the at least one anomaly detection unit is determined according to the probability value of each of the second flight time values.

5. The method according to claim 4, characterized in that The estimating, based on the flight time information, a first flight time value of a first conforming event on a response line corresponding to the at least one anomaly detection unit comprises: For each of the first coincidence events, a second flight time value is allocated to the first coincidence event according to the flight time information as the first flight time value of the first coincidence event.

6. The method according to claim 3, characterized in that The raw data includes third coincidence events, and each of the third coincidence events corresponds to a third flight time value.

7. The method according to claim 6, characterized in that Generating an initial image of the scanned object according to the original data includes: For each of the third coincident events in the original data, determining an ideal annihilation position of the third coincident event according to the corresponding third flight time value; The initial image is generated according to the ideal annihilation position of each of the third coincident events.

8. A positron emission tomography system, characterized in that: It includes an acquisition module, a count estimation module, a flight time information estimation module, a flight time value estimation module and a reconstruction module; The acquisition module is used to acquire raw data of the scanned object, wherein the raw data is obtained by detecting gamma photons by a plurality of detection units of a positron emission tomography detector, wherein the plurality of detection units include at least one anomaly detection unit; The count estimation module is used to estimate the count of the first coincidence event on the response line corresponding to the at least one anomaly detection unit according to the original data; The flight time information estimation module is used to estimate the flight time information related to the response line corresponding to the at least one anomaly detection unit based on the original data; The flight time value estimation module is used to estimate a first flight time value of a first conforming event on a response line corresponding to the at least one anomaly detection unit according to the flight time information; The reconstruction module generates a positron emission tomography image of the scanned object according to the original data and the first flight time value of the first coincident event.

9. A positron emission tomography apparatus, comprising a processor, wherein the processor is used to execute the positron emission tomography method according to any one of claims 1 to 7.

10. A computer-readable storage medium, wherein the storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the positron emission tomography method according to any one of claims 1 to 7.