PET image reconstruction method, device, medium and product with self-attenuation correction

Abstract

The present invention discloses a PET image reconstruction method, device, medium, and product with self-attenuation correction. The method includes: obtaining measurement data and a total tracer concentration corresponding to the measurement data, the measurement data being obtained using a time-of-flight positron emission tomography imaging method; alternatingly updating the concentration map and the attenuation sinusoidal map based on the measurement data, the total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and a sensitivity matrix until the iteration ends, thereby obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image; the concentration map update rules and the attenuation sinusoidal map update rules are both determined based on a minimization model, the minimization model being used to minimize the negative logarithmic maximum likelihood function of a measurement data model, the minimization model including a concentration constraint term, and the measurement data model being used to constrain the measurement data to decay exponentially. Embodiments of the present invention can reconstruct high-quality PET images through self-attenuation correction.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a PET image reconstruction method, device, medium and product with self-attenuation correction. Background Art

[0002] The physical mechanism of PET imaging (positron emission tomography) is that a radioactive isotope-labeled tracer enters the target subject and decays within the body, releasing positrons. These positrons instantly annihilate with surrounding electrons, producing a pair of gamma photons with energies of 511 keV, emitted in opposite directions at approximately 180 degrees. By measuring these paired gamma photons, the tracer's distribution within the body is reconstructed, and based on the varying tracer concentrations in different regions, the metabolic functional characteristics of the target's internal organs and tissues are determined. However, because gamma rays interact with substances in the body, the gamma photons measured by the detector have undergone attenuation. To obtain high-quality, quantitatively accurate PET images, attenuation correction of the measurement data is essential.

[0003] Currently, the most commonly used correction method is to scan with additional imaging equipment, estimate the attenuation coefficient in the body in advance, and then perform attenuation correction on the measured data based on the attenuation coefficient. As early as the 1980s, time-of-flight (TOF) technology was proposed. As the name suggests, this technology can measure the time difference between pairs of photons arriving at the detector. In theory, the location of the annihilation point can be completely determined by measuring only the time difference, but in fact, this technology is limited by the accuracy of the time resolution. Based on the TOF principle, in theory, the location of the annihilation point can be directly reversed by only collecting data through PET to achieve quantitative image reconstruction, but the existing technology is limited by the accuracy of the time resolution and cannot reconstruct high-quality PET images based on TOF. Summary of the Invention

[0004] The present invention provides a PET image reconstruction method, device, medium and product with self-attenuation correction to solve the problem that the existing technology cannot reconstruct high-quality PET images based on TOF.

[0005] According to one aspect of the present invention, a PET image reconstruction method with self-attenuation correction is provided, comprising:

[0006] Acquiring measurement data for reconstructing a PET image of a target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is acquired using a positron emission tomography imaging method based on time of flight;

[0007] Based on the measurement data, the total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and sensitivity matrix, alternately updating the concentration map and the attenuation sinusoidal map until the iteration ends, obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image;

[0008] Among them, the concentration map update rule and the attenuation sinusoidal map update rule are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

[0009] According to another aspect of the present invention, there is provided a PET image reconstruction apparatus with self-attenuation correction, comprising:

[0010] an acquisition module, configured to acquire measurement data for reconstructing a PET image of a target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is acquired using a positron emission tomography imaging method based on time of flight;

[0011] an image reconstruction module, configured to alternately update the concentration map and the attenuation sinusoidal map based on the measurement data, the total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and a sensitivity matrix until the iteration ends, thereby obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image;

[0012] Among them, the concentration map update rule and the attenuation sinusoidal map update rule are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

[0013] According to another aspect of the present invention, an electronic device is provided, comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the PET image reconstruction method with self-attenuation correction described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the PET image reconstruction method with self-attenuation correction described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned PET image reconstruction methods with self-attenuation correction.

[0019] The technical solution of the embodiment of the present invention, on the one hand, uses the exponential form of self-attenuation correction as a priori to determine the measurement data model, thereby improving the accuracy of the photon attenuation description; on the other hand, a minimization model is determined based on the negative log-likelihood function of the measurement data model. The concentration constraint term in the minimization model makes the alternating update process of the concentration map and the attenuation sinusoidal map more stable and easier to converge. Therefore, with the same number of iterations, a higher-precision concentration map and attenuation sinusoidal map can be obtained; on the other hand, the update process of the concentration map and the attenuation sinusoidal map does not depend on the selection of parameters. Therefore, there is no need to adjust the parameters during the update process of the concentration map and the attenuation sinusoidal map, which simplifies the image reconstruction process; finally, the reconstruction of the concentration map is completed through self-attenuation correction, which simplifies the scanning process of PET imaging and helps to improve the patient experience.

[0020] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0022] Figure 1 is a flow chart of a PET image reconstruction method with self-attenuation correction according to an embodiment of the present invention;

[0023] Figure 2 is another flow chart of a PET image reconstruction method with self-attenuation correction provided according to an embodiment of the present invention;

[0024] Figure 3 is a first true concentration map of the phantom provided according to an embodiment of the present invention;

[0025] Figure 4 is a first attenuation CT (computed tomography) image of the phantom provided according to an embodiment of the present invention;

[0026] Figure 5 is a first true attenuation sinusoid of the phantom provided according to an embodiment of the present invention;

[0027] Figure 6 is a first target concentration map of the phantom provided according to an embodiment of the present invention;

[0028] Figure 7 is a first target attenuation sinusoid of the phantom provided according to an embodiment of the present invention;

[0029] Figure 8 2 is a schematic diagram showing how the relative error of a concentration map varies with the number of iterations according to an embodiment of the present invention;

[0030] Figure 9 is a schematic diagram showing how the relative error of a decaying sinusoidal graph varies with the number of iterations according to an embodiment of the present invention;

[0031] Figure 10 2 is a schematic diagram showing how the relative error of data provided by an embodiment of the present invention changes with the number of iterations;

[0032] Figure 11 2. It is a schematic diagram showing how the function value of the minimization model changes with the number of iterations according to an embodiment of the present invention;

[0033] Figure 12 is a second true concentration map of the phantom provided according to an embodiment of the present invention;

[0034] Figure 13 is a second attenuation CT image of the phantom provided according to an embodiment of the present invention;

[0035] Figure 14 is a second true attenuation sinusoid of the phantom provided according to an embodiment of the present invention;

[0036] Figure 15 is a second target concentration map of the phantom provided according to an embodiment of the present invention;

[0037] Figure 16 is a third target concentration map of the phantom provided according to an embodiment of the present invention;

[0038] Figure 17 is a schematic structural diagram of a PET image reconstruction device with self-attenuation correction according to an embodiment of the present invention;

[0039] Figure 18 is another structural schematic diagram of a PET image reconstruction device with self-attenuation correction provided according to an embodiment of the present invention;

[0040] Figure 19 It is a structural schematic diagram of an electronic device for implementing the PET image reconstruction method with self-attenuation correction according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of the present invention.

[0042] It should be noted that the terms "object" and "target" in the description and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the numbers used in this way can be interchanged where appropriate so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0043] Figure 1 The flowchart of the PET image reconstruction method with self-attenuation correction provided by an embodiment of the present invention is applicable to the case of reconstructing PET images corresponding to measurement data acquired based on TOF-PET (positron emission tomography based on time of flight) scanning. The method can be performed by a PET image reconstruction device with self-attenuation correction. The PET image reconstruction device with self-attenuation correction can be implemented in the form of hardware and / or software. The PET image reconstruction device with self-attenuation correction can be configured in an electronic device. Figure 1 As shown, the method includes:

[0044] S110 , obtaining measurement data for reconstructing a PET image of a target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is obtained using a positron emission tomography imaging method based on time of flight.

[0045] Tracers, also called contrast agents, are positron-emitting radionuclides that are injected before a PET scan. The most widely used is 18F-fluorodeoxyglucose (18F-FDG), a positron-emitting form of glucose that participates in human metabolism.

[0046] A predetermined total concentration of tracer is injected into the target subject. After a set time, the PET imaging device is controlled to perform an imaging scan of the target subject using the TOF-PET method to obtain measurement data for the target subject at that total concentration. It is understood that different total concentrations correspond to different measurement data.

[0047] S120. Based on the measurement data, total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and sensitivity matrix, alternately update the concentration map and the attenuation sinusoidal map until the iteration ends, obtain a target concentration map and a target attenuation sinusoidal map, and use the target concentration map as the target PET image; wherein, the concentration map update rule and the attenuation sinusoidal map update rule are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

[0048] The concentration map update rule can be understood as the concentration map update formula; the attenuation sinusoidal map update rule can be understood as the attenuation sinusoidal map update formula.

[0049] Each element in the sensitivity matrix can be expressed as a ijt , which represents the probability that the photon generated in the j-th pixel block or voxel block is detected by the i-th response line in the t-th flight time window.

[0050] The measurement data model is used to simulate the attenuation form and attenuation results of the measurement data. The measurement data in the measurement data model of this embodiment is constrained to decay in an exponential form. Specifically: the first mean of the Poisson random variable corresponding to the number of photons collected by the t-th window of each response line is equal to the product of the second mean of the unattenuated measurement data of each response line in the t-th window and the attenuation sinusoid corresponding to each response line that decays in an exponential form; the second mean is equal to the sum of the measurement data corresponding to each pixel block under the t-th window, and the measurement data corresponding to each pixel block is equal to the product of the tracer concentration value in each pixel block and the probability that the photons generated by each pixel block are detected by the t-th window of each response line.

[0051] Measurement Data Model It can be expressed as:

[0052]

[0053] In this measurement data model, s i represents the line integral of the attenuation graph on the i-th response line, where i = 1,…,I represents different response lines (LOR); is m it The mean value, m it represents the Poisson random variable corresponding to the number of photons collected by the t-th time-of-flight window (TOF) of the i-th LOR. In addition, a ijt represents the probability that the photon generated in the jth pixel block or voxel block is detected by the i-th response line within the t-th flight time window, λ j represents the tracer concentration value in the jth pixel block or voxel block, J represents the total number of pixel blocks or voxel blocks; t = 1, ..., T represents different time-of-flight windows. Let It represents the mean of the measurement data without attenuation for the i-th response line in the t-th flight time window.

[0054] It should be noted that the attenuation of the measured data can be simply constrained to It can also be constrained to f∶[0,+∞)→[0,+∞) is a one-to-one mapping, as well as other exponential decay forms.

[0055] In order to facilitate the description of the technical solution, The vector formed by λ is called the concentration diagram, and the vector formed by s is called the attenuation sinusoidal diagram.

[0056] In TOF-PET, based on the above modeling, recovering the unknown tracer distribution while accounting for attenuation correction is equivalent to simultaneously solving for λ and s. The target concentration map is the target PET image, which can be used for clinical diagnosis and is the information of interest to physicians. s does not play a direct role in actual clinical diagnosis.

[0057] Let m it Represents the detection data corresponding to the i-th LOR under the t-th flight time window, and sets all random variables m it are independent of each other, then mathematically, the simultaneous reconstruction problem in TOF-PET can be equivalently converted into a minimization model, which includes multiple constraints and negative log-likelihood functions, such as:

[0058]

[0059] Where δ(·) is an indicator function, that is, if the constraint is satisfied, the function takes a value of zero, otherwise it takes an infinite value; 1 represents an all-one vector of length J; represents the sum of all components of λ, and N represents the total known concentration (total annihilation amount).

[0060] Specifically, the minimization model includes a negative log-likelihood term Non-negative concentration term (δ(λ≥0)), non-negative decaying sinusoidal term (δ(s≥0)), and concentration constraint term The concentration constraint ensures that the sum of the tracer concentrations corresponding to all concentration maps determined in each iteration is equal to the total concentration. In other words, the total concentration reconstructed in each iteration is required to remain consistent with the total tracer concentration injected into the target subject. Therefore, the concentration constraint can improve the speed and stability of iterative convergence.

[0061] In one embodiment, each component of the non-negative concentration map corresponding to the non-negative concentration map item is set to be greater than or equal to zero in a defined form, for example, directly defining λ≥0. Each component of the non-negative attenuation sinusoidal map corresponding to the non-negative attenuation sinusoidal map item is set to be greater than or equal to zero in a defined form, for example, directly defining s≥0. The concentration constraint term is directly configured as

[0062] It's understandable that directly updating both λ and s simultaneously isn't feasible. Therefore, this embodiment uses an alternating update method, alternating between λ and s. Alternating updates can be understood as maintaining the concentration map unchanged and updating the attenuation sinusoidal map; then maintaining the attenuation sinusoidal map unchanged and updating the concentration map; then maintaining the concentration map unchanged and updating the attenuation sinusoidal map again, and so on.

[0063] Based on formula (2), we can get the concentration map update rule, that is, the concentration map update formula. The details are as follows:

[0064]

[0065] Where k is the identifier of the current iteration round.

[0066] Let the initial concentration map λ 0 Let the initial decay sinusoidal graph s be a fully one vector. 0 The concentration map is updated based on formula (3) by setting the vector to zero and keeping the attenuation sinusoidal graph unchanged. After the concentration map is updated, the relative sizes of the components in the updated concentration map are adjusted based on the total concentration defined by the concentration constraint and the principle that the sum of the components in the concentration map remains unchanged, so as to update the concentration map again. The details are as follows:

[0067]

[0068] The updated concentration map is used as the current concentration map. Keeping the current concentration map unchanged, the attenuation sinusoidal map is updated based on the following formula:

[0069]

[0070] After the attenuation sinusoid is updated, based on the principle that each element in the attenuation sinusoid is non-negative, each negative element in the updated attenuation sinusoid is set to zero to update the attenuation sinusoid again, as follows:

[0071]

[0072] Similarly, the concentration map and attenuation sinusoidal map are updated alternately until the updated concentration map, attenuation sinusoidal map, or the number of iterations reaches the corresponding termination condition, and the iteration ends; the concentration map and attenuation sinusoidal map determined in the last iteration round are used as the target concentration map and target attenuation sinusoidal map.

[0073] The termination condition for the concentration map is that the first image evaluation index of the concentration map determined in adjacent iterations is within a first set range; the termination condition for the attenuation sinusoidal map is that the second image evaluation index of the attenuation sinusoidal map determined in adjacent iterations is within a second set range. The first image evaluation index and the second image evaluation index can be the same or different; the first set range and the second set range can be the same or different, and their sizes can be determined by the user based on the quality requirements of the concentration map and the attenuation sinusoidal map. In one embodiment, the first image evaluation index and the second image evaluation index are the same, and both are relative error or structural similarity, etc.

[0074] The termination condition corresponding to the number of iterations is the set number of iterations, such as 50, 500, 1000, 10000, etc.

[0075] During the alternating update process of the concentration map and the attenuation sinusoidal map, if the termination condition corresponding to the concentration map is used to determine whether to continue the iteration, then if the concentration map determined in any iterative round does not meet the corresponding termination condition, the concentration map is kept unchanged and the attenuation sinusoidal map is updated, and then the next iterative round is entered; if the concentration map determined in any iterative round meets the corresponding termination condition, the iteration can be terminated, and the concentration map determined in the iterative round is used as the target concentration map, and the attenuation sinusoidal map determined in the iterative round is used as the target attenuation sinusoidal map.

[0076] Because the concentration map update process is related to the changes in the attenuation sinogram, once the attenuation sinogram is determined, the concentration map based on it is also determined. Based on this, during the alternating update process of the concentration map and the attenuation sinogram, if the termination condition corresponding to the attenuation sinogram is used to determine whether to continue the iteration, then if the attenuation sinogram determined in any iteration round does not meet the corresponding termination condition, the next iteration round will be entered; if the attenuation sinogram determined in any iteration round meets the corresponding termination condition, the iteration can be terminated, and the concentration map determined in that iteration round will be used as the target concentration map, and the attenuation sinogram determined in that iteration round will be used as the target attenuation sinogram.

[0077] During the alternating update process of the concentration map and the attenuation sinusoidal map, if the termination condition corresponding to the number of iterations is used to determine whether the iteration continues, when the number of iterations corresponding to any iteration round reaches the maximum number of iterations, the attenuation sinusoidal map determined in the iteration round is used as the target attenuation sinusoidal map, and the concentration map determined in the iteration round is used as the target concentration map, and then the iteration is stopped; or the iteration is stopped directly, at which time the previous iteration round is the last iteration round, so the attenuation sinusoidal map determined in the previous iteration round is used as the target attenuation sinusoidal map, and the concentration map determined in the previous iteration round is used as the target concentration map.

[0078] It should be noted that during the concentration map update process, the updates of the concentration maps corresponding to each pixel block or voxel block do not affect each other. Therefore, the concentration maps corresponding to different pixel blocks or voxel blocks can be updated serially or in parallel. Similarly, the updates of the attenuation sinusoidal graphs corresponding to each response line do not affect each other. Therefore, the attenuation sinusoidal graphs corresponding to different response lines can be updated serially or in parallel. In addition, in the initial iteration round, the attenuation sinusoidal graph can be kept unchanged at the starting attenuation sinusoidal graph, the concentration map can be updated, and then the updated attenuation sinusoidal graph can be updated while keeping the updated concentration map unchanged. Alternatively, the concentration map can be kept unchanged at the starting concentration map, the attenuation sinusoidal graph can be updated, and then the updated attenuation sinusoidal graph can be updated while keeping the updated attenuation sinusoidal graph unchanged.

[0079] In one embodiment, after the target concentration map and the target attenuation sinusoidal map are determined, the target concentration map is displayed in a visualization interface. The target concentration map is a PET image, so that the user can view the required PET image.

[0080] The technical solution provided by the embodiment of the present invention, on the one hand, uses the exponential form of self-attenuation correction as a priori to determine the measurement data model, thereby improving the accuracy of the photon attenuation description; on the other hand, a minimization model is determined based on the negative log-likelihood function of the measurement data model. The concentration constraint term in the minimization model makes the alternating update process of the concentration map and the attenuation sinusoidal map more stable and easier to converge. Therefore, with the same number of iterations, a higher-precision concentration map and attenuation sinusoidal map can be obtained; on the other hand, the update process of the concentration map and the attenuation sinusoidal map does not depend on the selection of parameters. Therefore, no parameter adjustment is required during the update process of the concentration map and the attenuation sinusoidal map, which simplifies the image reconstruction process; finally, the reconstruction of the concentration map is completed through self-attenuation correction, which simplifies the scanning process of PET imaging and helps to improve the patient experience.

[0081] Figure 2 This is another flow chart of the PET image reconstruction method with self-attenuation correction provided by an embodiment of the present invention. This embodiment adds a termination condition configuration step based on the previous embodiment. Figure 2 As shown, the method includes:

[0082] S210. In response to the confirmation operation of the configuration parameters of the parameter configuration interface, the configuration parameters are used as corresponding termination conditions. The configuration parameters are at least one of the first boundary condition, the second boundary condition and the maximum number of iterations. The first boundary condition corresponds to the first image evaluation index of the concentration graph, and the second boundary condition corresponds to the second image evaluation index of the attenuation sinusoidal graph.

[0083] In this embodiment, the termination condition used in the PET image reconstruction process is configured as a modifiable item. That is, the user can use the default termination condition to determine the timing of iterative interruption in the PET image reconstruction process, or can configure the termination condition according to actual conditions.

[0084] If the user wishes to configure their own termination conditions, they can open the parameter configuration interface and configure the corresponding parameters there, such as the first boundary condition corresponding to the first image evaluation metric of the concentration map, the second boundary condition corresponding to the second image evaluation metric of the attenuation sinusoidal map, or the maximum number of iterations. After configuring the parameters, they can click or touch the Confirm option in the parameter configuration interface. In response to this confirmation, the processor will use the configured parameters in the configuration interface as the corresponding termination conditions.

[0085] Regarding the first image evaluation index and the second image evaluation index, please refer to the aforementioned embodiment, which will not be described in detail in this embodiment. The first boundary condition corresponds to the first set range. For example, if it is the upper boundary of the first set range, the lower boundary of the first set range defaults to 0. The second boundary condition corresponds to the second set range. For example, if the upper boundary of the second set range is the second boundary condition, the lower boundary of the second set range defaults to 0.

[0086] S220 , obtaining measurement data for reconstructing a PET image of the target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is obtained using a positron emission tomography imaging method based on time of flight.

[0087] S230. Based on the measurement data, total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and sensitivity matrix, the concentration map and the attenuation sinusoidal map are alternately updated until the iteration ends, thereby obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image; wherein the concentration map update rules and the attenuation sinusoidal map update rules are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

[0088] In one embodiment, the first evaluation metric is a first relative error, and the first boundary condition includes a first relative error threshold. During the alternating update of the concentration map and the attenuation sinusoidal map, if it is detected that the relative error between the concentration maps determined in adjacent iterations is less than or equal to the first relative error threshold, the iteration is terminated. The concentration map determined in the current iteration is used as the target concentration map, and the attenuation sinusoidal map determined in the current iteration is used as the target attenuation sinusoidal map.

[0089] In one embodiment, the second evaluation metric is a second relative error, and the second boundary condition is a second relative error threshold. During the alternating update process of the concentration map and the attenuation sinusoidal map, if it is detected that the relative error between the attenuation sinusoidal maps determined in adjacent iterations is less than or equal to the second relative error threshold, the iteration is terminated. The concentration map determined in the current iteration is used as the target concentration map, and the attenuation sinusoidal map determined in the current iteration is used as the target attenuation sinusoidal map.

[0090] In one embodiment, the termination condition is a combination of a first relative error threshold, a second relative error threshold, and a maximum number of iterations. During the alternating update of the concentration map and the attenuation sinusoidal map, if the relative error between the concentration maps determined in adjacent iterations is less than or equal to the first relative error threshold, and if the relative error between the attenuation sinusoidal maps determined in adjacent iterations is greater than or equal to the second relative error, the iteration is terminated.

[0091] In the embodiment of the present invention, the parameter configuration function corresponding to the parameter configuration interface allows the user to set reasonable termination conditions based on actual needs, thereby improving the flexibility and universality of PET image reconstruction.

[0092] In one embodiment, the PET image reconstruction process is refined by taking the noise-free data of the reconstructed body as an example. The real image is determined by simulating the phantom. The real image corresponding to the phantom is distributed in an area of ​​size [-15cm, 15cm] × [-15cm, 15cm], with a pixel size of 64 × 64, such as Figure 3 The first true concentration map shown is Figure 4 The first attenuation CT image shown and Figure 5 The first real attenuation sinusoidal graph is shown. The grayscale window of the first real attenuation sinusoidal graph is [0, 0.115] cm -1 .

[0093] The first measurement data of the phantom is acquired by parallel beam scanning. Specifically, 64 scanning angles are evenly distributed from 0 to 180 degrees, and 64 detectors are evenly distributed in the range of [-15 cm, 15 cm]. The TOF time resolution considered is approximately 600 ps, ​​and a total of 10 TOF windows with a window width of 4.5 cm are set. The discrete X-ray transformation and the real image are substituted into the imaging model to simulate the generation of noise-free first measurement data, and then the PET image reconstruction method with self-attenuation correction described in the above embodiment is used to reconstruct the noise-free first measurement data to obtain a first target concentration map (see Figure 6 ) and the first target attenuation sinusoid (see Figure 7 ). It is obvious that the first target concentration map ( Figure 6 ) and the first true concentration map ( Figure 3 ) is almost the same, the first target attenuation sinusoidal diagram ( Figure 7 ) and the first true decay sinusoidal graph ( Figure 5 ), which shows that the method described in the embodiment of the present invention can accurately reconstruct quantitative PET images.

[0094] In order to quantitatively measure the performance of the algorithm corresponding to the embodiment of the present invention, the number of iterations is set as large as possible, for example, 10 million times, and a curve of the relative error of the concentration map versus the number of iterations is plotted ( Figure 8 ), the curve of the relative error of the attenuation sinusoidal graph versus the number of iterations ( Figure 9 ), the curve of the relative error of the data versus the number of iterations ( Figure 10 ) and the graph of the objective function value versus the number of iterations ( Figure 11 ). The objective function value can be understood as the function value of the minimization model when the concentration map and attenuation sinusoidal map determined by each iteration are input into the minimization model; the data can be understood as the difference between the measured data inferred based on the concentration map and attenuation sinusoidal map determined by each iteration and the actual measured data. Figure 11 It can be seen from the graph of the relative error of the objective function value versus the number of iterations that the method described in the embodiment of the present invention is convergent on the target value, and the curve of the relative error of the concentration graph versus the number of iterations, the curve of the relative error of the attenuation sinusoidal graph versus the number of iterations, and the curve of the relative error of the data versus the number of iterations also fully demonstrate that the method described in the embodiment of the present invention has high stability and convergence.

[0095] In one embodiment, the true concentration map is determined by simulating a phantom, and distributed within an area of ​​[-15 cm, 15 cm] × [-15 cm, 15 cm], with a pixel size of 128 × 128, such as Figure 12 The second true concentration diagram is shown as Figure 13The second real attenuation CT image shown and Figure 14 The second true decay sinusoid is shown. The scanning geometry used is the same as before. Specifically, 128 scan angles are set, evenly distributed between 0 and 180 degrees; there are 128 detectors, equally spaced within the range [-15 cm, 15 cm]. The TOF temporal resolution used is still 600 ps, ​​with a total of 10 TOF windows of 4.5 cm width. The process for generating the noise-free second measurement data and implementing the algorithm is the same as before.

[0096] The PET image reconstruction method with self-attenuation correction described in the above embodiment is used to determine the target concentration map corresponding to the second measurement data, that is, the second target concentration map (such as Figure 15 ).

[0097] In order to quantitatively compare the quality of the reconstructed PET images (target concentration maps), the following indicators are used to measure the quality of the PET images (target concentration maps). Given two images, the vectors are and u * , their normalized root mean square error is defined as:

[0098]

[0099] The peak signal-to-noise ratio is defined as:

[0100]

[0101] Where I is the total number of image pixels.

[0102] Structural similarity is defined as:

[0103]

[0104] Among them, x i and y i Corresponding to and u * The i-th 8×8 window (image block) is represented by M, and M represents the total number of windows.

[0105] The local similarity index is defined as:

[0106]

[0107] in, and Represent the mean and variance of x and y respectively, and c1=c2=0.05.

[0108] The second target concentration map determined by the scheme described in this embodiment ( Figure 15 ) and the second true concentration map ( Figure 12) is 0.00113, the peak signal-to-noise ratio is 76.02, and the result of subtracting the structural similarity from one is 5.97e-6. Thus, the method described in the embodiment of the present invention reconstructs the target PET image with high accuracy and image quality.

[0109] In one embodiment, the true concentration map is determined by simulating a phantom, and distributed within an area of ​​[-15cm, 15cm] × [-15cm, 15cm], with a pixel size of 128 × 128, such as Figure 12 The second true concentration map is shown. The scanning geometry used is the same as before. Specifically, 128 scanning angles are set, evenly distributed between 0 and 180 degrees. There are 128 detectors, equidistantly spaced within the range [-15 cm, 15 cm]. The TOF temporal resolution used is still 600 ps, ​​and a total of 10 TOF windows with a width of 4.5 cm are set. The process of generating the noise-free second measurement data and implementing the algorithm are the same as before.

[0110] The second measurement data is put into a Poisson generator to obtain measurement data containing Poisson noise, namely the third measurement data. In the experiment, the noise level is set to 7.24dB. The number of iterations is set to 50 steps, and the above-mentioned PET image reconstruction method with self-attenuation correction is used to reconstruct the third measurement data to obtain the third target concentration map ( Figure 16 ), it was found that the normalized root mean square error between the third target concentration map and the second real image (12) was 0.2520, the peak signal-to-noise ratio was 29.02, and the structural similarity was 0.8996. This shows that even for measurement data containing noise, the method of the present invention can still reconstruct a high-quality target concentration map, so the method of the present invention has high robustness.

[0111] Figure 17 Schematic diagram of the structure of a PET image reconstruction device with self-attenuation correction provided by an embodiment of the present invention. Figure 17 As shown, the device includes:

[0112] an acquisition module 310, configured to acquire measurement data for reconstructing a PET image of a target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is acquired using a positron emission tomography imaging method based on time of flight;

[0113] an image reconstruction module 320 for alternately updating the concentration map and the attenuation sinusoidal map based on the measurement data, the total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and a sensitivity matrix until the iteration ends, thereby obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image;

[0114] Among them, the concentration map update rule and the attenuation sinusoidal map update rule are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

[0115] In one embodiment, the minimization model includes a negative log-likelihood function term, a non-negative concentration graph term, a non-negative decay sinusoidal graph term, and a concentration constraint term.

[0116] In one embodiment, when the concentration map is greater than or equal to zero, the function value of the non-negative concentration map term is equal to zero;

[0117] In the case where the attenuation sinusoidal graph is greater than or equal to zero, the function value of the non-negative attenuation sinusoidal graph term is equal to zero;

[0118] When the sum of the tracer concentrations corresponding to all the concentration maps is equal to the total concentration, the function value of the concentration constraint term is equal to zero, otherwise it is infinite.

[0119] In one embodiment, during the process of alternatingly updating the concentration map and the attenuation sinusoidal map, the image reconstruction module is specifically configured to:

[0120] While keeping the attenuation sinusoidal graph unchanged and completing the concentration map update, adjusting the relative sizes of the components in the updated concentration map based on the principle that the sum of the components in the concentration map remains unchanged, so as to update the concentration map again;

[0121] While keeping the concentration map unchanged and completing the update of the attenuation sinusoidal map, based on the principle that each element in the attenuation sinusoidal map is non-negative, each negative element in the updated attenuation sinusoidal map is set to zero to update the attenuation sinusoidal map again.

[0122] In one embodiment, the measurement data model is configured as follows:

[0123] The measurement data model is configured as follows:

[0124] The first mean of the Poisson random variable corresponding to the number of photons collected in the t-th window of each response line is equal to the product of the second mean of the unattenuated measurement data of each response line in the t-th window and the exponentially decaying attenuation sinogram corresponding to each response line;

[0125] The second mean is equal to the sum of the measurement data corresponding to each pixel block under the t-th window, and the measurement data corresponding to each pixel block is equal to the product of the tracer concentration value in each pixel block and the probability that the photon generated by each pixel block is detected by the t-th window of each response line.

[0126] In one embodiment, Figure 18 As shown, the device further includes a configuration module 33003, and the configuration module 300 is used to:

[0127] In response to a confirmation operation on a configuration parameter of a parameter configuration interface, taking the configuration parameter as a corresponding termination condition;

[0128] The configuration parameter is at least one of a first boundary condition, a second boundary condition, and a maximum number of iterations, the first boundary condition corresponds to a first image evaluation index of the concentration map, and the second boundary condition corresponds to a second image evaluation index of the attenuation sinusoidal map.

[0129] In one embodiment, the image reconstruction module is specifically configured to:

[0130] Alternately updating the concentration map and the attenuation sinusoidal graph until the updated concentration map, the attenuation sinusoidal graph, or the number of iterations reaches a corresponding termination condition, thereby terminating the iteration;

[0131] The concentration map and attenuation sinusoidal map determined in the last iteration round are used as the target concentration map and target attenuation sinusoidal map.

[0132] In one embodiment, the device further comprises a display module, wherein the display module is configured to:

[0133] The target concentration graph is displayed in a visual interface.

[0134] The technical solution provided by the embodiment of the present invention, on the one hand, uses the exponential form of self-attenuation correction as a priori to determine the measurement data model, thereby improving the accuracy of the photon attenuation description; on the other hand, a minimization model is determined based on the negative log-likelihood function of the measurement data model. The concentration constraint term in the minimization model makes the alternating update process of the concentration map and the attenuation sinusoidal map more stable and easier to converge. Therefore, with the same number of iterations, a higher-precision concentration map and attenuation sinusoidal map can be obtained; on the other hand, the update process of the concentration map and the attenuation sinusoidal map does not depend on the selection of parameters. Therefore, no parameter adjustment is required during the update process of the concentration map and the attenuation sinusoidal map, which simplifies the image reconstruction process; finally, the reconstruction of the concentration map is completed through self-attenuation correction, which simplifies the scanning process of PET imaging and helps to improve the patient experience.

[0135] The PET image reconstruction device with self-attenuation correction provided by the embodiment of the present invention can execute the PET image reconstruction method with self-attenuation correction provided by any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0136] Figure 19 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0137] like Figure 19 As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., which is communicatively connected to the at least one processor 11. The memory stores a computer program that can be executed by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the RAM 13. The processor 11, ROM 12, and RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0138] Multiple components in the electronic device 10 are connected to the I / O interface 15, including an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0139] The processor 11 may be any general-purpose and / or specialized processing component with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various specialized artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the PET image reconstruction method with self-attenuation correction.

[0140] In some embodiments, the PET image reconstruction method with self-attenuation correction can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the PET image reconstruction method with self-attenuation correction described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to perform the PET image reconstruction method with self-attenuation correction in any other suitable manner (e.g., by means of firmware).

[0141] Various embodiments of the systems and techniques described herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), system-on-chip systems (SOCs), programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include being implemented in one or more computer programs that are executable and / or interpreted on a programmable system that includes at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0142] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowcharts and / or block diagrams are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package, partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0143] In the context of the present invention, computer-readable storage media can be tangible media that can contain or store a computer program for use with an instruction execution system, device or equipment or used in combination with an instruction execution system, device or equipment. Computer-readable storage media can include but are not limited to electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, computer-readable storage media can be machine-readable signal media. More specific examples of machine-readable storage media can include electrical connections based on one or more lines, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0144] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0145] The systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0146] A computing system may include clients and servers. The clients and servers are typically remote from each other and typically interact via a communication network. This client-server relationship arises through computer programs running on the respective computers, creating a client-server relationship. The server may be a cloud server, also known as a cloud computing server or cloud host. This server is a hosting product within the cloud computing service ecosystem that addresses the management difficulties and limited scalability of traditional physical hosting and VPS services.

[0147] An embodiment of the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the PET image reconstruction method with self-attenuation correction as provided in any embodiment of the present application.

[0148] The computer program product may be implemented by writing computer program code for performing the operations of the present invention in one or more programming languages, or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0149] It should be understood that the various forms of the processes shown above can be used to reorder, add, or delete steps. For example, the steps described in the present invention can be performed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved. This is not limited herein.

[0150] The above specific embodiments do not limit the scope of protection of the present invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may be made based on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention are intended to be included within the scope of protection of the present invention.

Claims

1. A PET image reconstruction method with self-attenuation correction, characterized in that include: Acquiring measurement data for reconstructing a PET image of a target object and a total tracer concentration corresponding to the measurement data, wherein the measurement data is acquired using a positron emission tomography imaging method based on time of flight; Based on the measurement data, the total concentration, and pre-stored concentration map update rules, attenuation sinusoidal map update rules, and sensitivity matrix, alternately updating the concentration map and the attenuation sinusoidal map until the iteration ends, obtaining a target concentration map and a target attenuation sinusoidal map, and using the target concentration map as a target PET image; Among them, the concentration map update rule and the attenuation sinusoidal map update rule are both determined based on a minimization model, the minimization model is used to minimize the negative logarithmic maximum likelihood function of the measurement data model, the minimization model includes a concentration constraint term, the concentration constraint term is used to make the sum of the tracer concentrations corresponding to all concentration maps generated in each iteration equal to the total concentration, and the measurement data model is used to constrain the measurement data to decay exponentially.

2. The method according to claim 1, characterized in that The minimization model includes a negative logarithmic maximum likelihood term, a non-negative concentration graph term, a non-negative attenuated sinusoidal graph term and a concentration constraint term.

3. The method according to claim 2, characterized in that In the case where the concentration map is greater than or equal to zero, the function value of the non-negative concentration map item is equal to zero; In the case where the attenuation sinusoidal graph is greater than or equal to zero, the function value of the non-negative attenuation sinusoidal graph term is equal to zero; When the sum of the tracer concentrations corresponding to all the concentration maps is equal to the total concentration, the function value of the concentration constraint term is equal to zero, otherwise it is infinite.

4. The method according to claim 1, wherein During the alternating update of the concentration map and the attenuation sinusoidal map: While keeping the attenuation sinusoidal graph unchanged and completing the concentration map update, adjusting the relative sizes of the components in the updated concentration map based on the principle that the sum of the components in the concentration map remains unchanged, so as to update the concentration map again; While keeping the concentration map unchanged and completing the update of the attenuation sinusoidal map, based on the principle that each element in the attenuation sinusoidal map is non-negative, each negative element in the updated attenuation sinusoidal map is set to zero to update the attenuation sinusoidal map again.

5. The method according to claim 1, wherein The measurement data model is configured as follows: The first mean of the Poisson random variable corresponding to the number of photons collected in the t-th window of each response line is equal to the product of the second mean of the unattenuated measurement data of each response line in the t-th window and the exponentially decaying attenuation sinogram corresponding to each response line; The second mean is equal to the sum of the measurement data corresponding to each pixel block under the t-th window, and the measurement data corresponding to each pixel block is equal to the product of the tracer concentration value in each pixel block and the probability that the photon generated by each pixel block is detected by the t-th window of each response line.

6. The method according to claim 1, characterized in that Before acquiring measurement data for reconstructing a PET image of the target object and a total tracer concentration corresponding to the measurement data, the method further includes: In response to a confirmation operation on a configuration parameter of the parameter configuration interface, taking the configuration parameter as a corresponding termination condition; The configuration parameter is at least one of a first boundary condition, a second boundary condition, and a maximum number of iterations, the first boundary condition corresponds to a first image evaluation index of the concentration map, and the second boundary condition corresponds to a second image evaluation index of the attenuation sinusoidal map.

7. The method according to claim 1, characterized in that The alternate updating of the concentration map and the attenuation sinusoidal map until the iteration ends to obtain the target concentration map and the target attenuation sinusoidal map includes: Alternately updating the concentration map and the attenuation sinusoidal graph until the updated concentration map, the attenuation sinusoidal graph, or the number of iterations reaches a corresponding termination condition, thereby terminating the iteration; The concentration map and attenuation sinusoidal map determined in the last iteration round are used as the target concentration map and target attenuation sinusoidal map.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so as to enable the at least one processor to perform the PET image reconstruction method with self-attenuation correction according to any one of claims 1 to 7.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the PET image reconstruction method with self-attenuation correction according to any one of claims 1 to 7 when executed.

10. A computer program product, characterized in that The computer program product comprises a computer program, which, when executed by a processor, implements the PET image reconstruction method with self-attenuation correction according to any one of claims 1 to 7.

Images