Compton camera image reconstruction method, device, computer equipment and storage medium

Abstract

This application relates to the field of medical imaging technology and discloses a Compton camera image reconstruction method, apparatus, computer equipment, and storage medium. The method comprises obtaining a current weight distribution and a reference weight distribution corresponding to an imaging plane; simulating a real radioactive source based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane; and updating the elements in the current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set for image reconstruction of the Compton camera. The elements in the current source set are used to characterize the coordinates of the radioactive source corresponding to a Compton event occurring within the Compton camera. Thus, an image reconstruction method is implemented that can improve the imaging quality of a Compton camera, enabling the reconstructed image to more accurately reflect the actual distribution of the radioactive source, thereby providing a reliable basis for real-time monitoring of boron concentration during BNCT treatment.

Description

Technical Field

[0001] The present application relates to the field of medical imaging technology, and in particular to an image reconstruction method, apparatus, computer equipment, and storage medium for a Compton camera. Background Art

[0002] Boron Neutron Capture Therapy (BNCT) is an advanced cancer treatment technology capable of precisely irradiating tumor cells. However, the accuracy of real-time boron dose monitoring has become a key challenge hindering its further development. With technological advancements, Compton cameras, with their electronic collimation properties and 3D imaging advantages, have shown great potential among various methods for real-time boron concentration monitoring.

[0003] Although Compton cameras have many advantages in theory, there are still certain limitations in the current imaging reconstruction methods of related technologies. There is a need for an image reconstruction method that can significantly improve the imaging quality. Summary of the Invention

[0004] This application aims to solve, at least to some extent, one of the technical problems in the related art. To this end, this application proposes a method, apparatus, computer device, and storage medium for image reconstruction using a Compton camera. The main technical solutions adopted in this application include:

[0005] In a first aspect, an embodiment of the present application provides an image reconstruction method for a Compton camera, the method comprising: obtaining a current weight distribution and a reference weight distribution corresponding to an imaging plane; performing a real radioactive source simulation based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane; updating elements in a current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set to perform image reconstruction on the Compton camera; wherein the elements in the current source set are used to represent the coordinates of the radioactive source corresponding to a Compton event occurring in the Compton camera.

[0006] Optionally, the elements in the current source set are updated according to the simulated weight distribution and the reference weight distribution to obtain an updated source set, including: determining the current difference data between the simulated weight distribution and the reference weight distribution; updating the elements in the current source set according to the current difference data to obtain an updated source set.

[0007] Optionally, a real radiation source simulation is performed based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane, including: generating an inversion circle with a radius of a specified length in the imaging plane based on the current weight distribution, and obtaining the simulated weight value of each pixel square in the imaging plane as the simulated weight distribution corresponding to the imaging plane.

[0008] Optionally, the elements in the current source set are updated according to the current difference data to obtain an updated source set, including: determining a first probability density of the random element and a second probability density of any element in the current source set; wherein the random element is used to characterize a random point on the reconstructed cone of the Compton event corresponding to any element; adjusting the first probability density and the second probability density based on the current difference data respectively to obtain a first adjusted density of the random element and a second adjusted density of any element; updating the elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set.

[0009] Optionally, the current difference data includes the current first-order error distribution and the current second-order error distribution; based on the current difference data, the first probability density and the second probability density are adjusted respectively to obtain the first adjusted density of the random element and the second adjusted density of any element, including: using the current first-order error distribution to adjust the second probability density to obtain the second adjusted density of any element; using the current second-order error distribution to adjust the first probability density to obtain the first adjusted density of the random element.

[0010] Optionally, the elements in the current source set are updated according to the first adjusted density and the second adjusted density to obtain an updated source set, including: if the first adjusted density is greater than or equal to the second adjusted density, replacing any element with a random element to obtain an updated source set; if the first adjusted density is less than the second adjusted density, discarding the random element and retaining any element to obtain an updated source set.

[0011] Optionally, the simulated weight distribution and the reference weight distribution are obtained by normalization processing respectively, and the current difference data includes the current first-order error distribution and the current second-order error distribution; the current first-order error distribution and the current second-order error distribution are determined in the following manner: the current error weight distribution is subtracted from the simulated weight distribution by the reference weight distribution to obtain the current error weight distribution as the current first-order error distribution; each current error coefficient in the current error weight distribution is squared to obtain the current second-order error distribution.

[0012] Optionally, the specified length is determined by: determining an average scattering angle corresponding to a Compton event occurring in a Compton camera, and a distance between a scattering detector of the Compton camera and an imaging plane; and determining an average radius of a projected cross-section of a reconstructed cone of the Compton event on the imaging plane based on the average scattering angle and the distance as the specified length.

[0013] Optionally, the reference weight distribution is used to describe the first weight value of each pixel square in the imaging plane obtained by the analytical reconstruction algorithm, and the current weight distribution is the second weight value of each pixel square in the imaging plane in the current round of the iterative reconstruction algorithm; if the current round is the first round, the current weight distribution is obtained by the random source set algorithm; if the current round is not the first round, the current weight distribution is obtained by weight matching the updated source set obtained in the previous round of the current round with the pixels in the imaging plane.

[0014] In a second aspect, an embodiment of the present application provides an image reconstruction device for a Compton camera, the device comprising: a weight distribution acquisition module for acquiring a current weight distribution and a reference weight distribution corresponding to an imaging plane; a simulation weight determination module for simulating a real radioactive source based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane; a current source set update module for updating elements in a current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set for image reconstruction of the Compton camera; wherein the elements in the current source set are used to characterize the coordinates of the radioactive source corresponding to a Compton event occurring in the Compton camera.

[0015] In a third aspect, the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of any of the above methods when executing the computer program.

[0016] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of any of the above methods when the computer program is executed by a processor.

[0017] In a fifth aspect, the present invention provides a computer program product, comprising a computer program, which implements the steps of any of the above methods when executed by a processor.

[0018] In the above embodiment, by obtaining the current weight distribution corresponding to the imaging plane and performing a real radioactive source simulation operation based on it, a simulated weight distribution of the possible locations of the real radioactive source in the fitted imaging area is obtained. Subsequently, the simulated weight distribution and the reference weight distribution are used to locate areas with large weight deviations, and the elements in the current source set are further updated. The elements in the current source set are used to characterize the coordinates of the radioactive source corresponding to the Compton event occurring in the Compton camera. Finally, the Compton camera image is reconstructed based on the updated source set, so that the reconstructed image more accurately reflects the actual distribution of the radioactive source, effectively improving the image reconstruction quality and accuracy of the Compton camera, and providing a more reliable basis for real-time monitoring of boron concentration during BNCT treatment, thereby improving the safety and effectiveness of BNCT treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the specific implementation methods of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the specific implementation methods or the description of the prior art. Obviously, the drawings described below are some implementation methods of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1a This is a diagram of the imaging principle of a Compton camera provided according to an embodiment of the present application;

[0021] Figure 1b This is a flowchart of an image reconstruction method for a Compton camera according to one embodiment of the present application;

[0022] Figure 1c This is an example diagram of the dimensions of a plane source provided according to one embodiment of the present application;

[0023] Figure 1d A diagram showing the structure and dimensions of a Compton camera according to one embodiment of the present application;

[0024] Figure 1e This is an image obtained by a direct back-projection algorithm according to one embodiment of the present application;

[0025] Figure 1f This is an imaging diagram of a random source set algorithm provided according to one embodiment of the present application;

[0026] Figure 1g This is an imaging diagram of an image reconstruction method of a Compton camera provided according to one embodiment of the present application;

[0027] Figure 1h This is an example diagram of plane source dimensions provided according to yet another embodiment of the present application;

[0028] Figure 1i This is an imaging diagram of a direct back-projection algorithm provided according to another embodiment of the present application;

[0029] Figure 1j This is an imaging diagram of a random source set algorithm provided according to another embodiment of the present application;

[0030] Figure 1k This is an imaging diagram of an image reconstruction method of a Compton camera provided according to another embodiment of the present application;

[0031] Figure 2 A flowchart of determining an updated source set according to one embodiment of the present application;

[0032] Figure 3 A flowchart for determining adjusted density according to one embodiment of the present application;

[0033] Figure 4 A flowchart of determining error distribution according to one embodiment of the present application;

[0034] Figure 5 A flowchart for determining a specified length according to one embodiment of the present application;

[0035] Figure 6 is a structural block diagram of an image reconstruction device for a Compton camera according to one embodiment of the present application;

[0036] Figure 7 The figure is a diagram of the internal structure of a computer device according to one embodiment of the present application. DETAILED DESCRIPTION

[0037] To make the purpose, technical solutions, and advantages of the embodiments of the present application more clear, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0038] Boron Neutron Capture Therapy (BNCT) is a cutting-edge and highly effective cancer treatment technology. Its core advantage lies in its ability to precisely target tumor cells with radiation, effectively killing diseased cells. Specifically, during treatment, a compound containing boron-10 is injected into the patient's body. This compound selectively accumulates in cancer cells while being less distributed in normal tissue. Some time after administration, when the boron drug reaches a sufficient concentration in the tumor tissue, a neutron beam is used to irradiate the tumor site. The neutrons react with the boron-10 in the tumor cells, producing alpha particles and lithium-7 ions. These particles deposit energy within the cells, killing the tumor cells and achieving the therapeutic effect.

[0039] However, despite the remarkable therapeutic effects of BNCT, accurate and real-time monitoring of the local boron dose, i.e., the radiation source, remains a key challenge hindering its further development. Currently, seven main methods are used to measure boron concentration during BNCT, but each has limitations. For example, while positron emission tomography (PET) can provide reference values ​​for boron concentration, it lacks real-time imaging. Prompt gamma-ray spectrometry, while capable of real-time imaging, presents significant challenges when measuring non-uniform boron concentrations. Several other methods also have their own advantages and disadvantages, and most are offline measurement methods. Therefore, the development of novel online boron concentration monitoring technologies is urgently needed. The Compton camera is an imaging technology based on an electron collimator. It utilizes the Compton scattering dynamics of photons to locate the location of gamma-ray generation. Specifically, the incident direction is determined based on the energy of the recoil electrons generated by the incident gamma-rays. This is equivalent to using a conical collimator to define the direction of the radiation source. It offers advantages such as high imaging efficiency, high sensitivity, and the ability to achieve three-dimensional imaging from a single angle.

[0040] Based on this, during the neutron irradiation process of boron neutron capture therapy, a Compton camera can be used to detect the gamma rays generated by the reaction of neutrons and boron-10 nuclei. According to the energy spectrum and distribution information of the gamma rays, the Compton scattering principle is used to invert the concentration distribution of boron in the target area through data processing and analysis. Specifically, the typical structure of a Compton camera can be composed of two layers of detectors, which are placed in parallel. The layer of detectors close to the space to be detected is called a scattering layer, and the other layer of detectors is called an absorption layer. And the two-layer detector structure mostly uses an array structure. For example, the imaging principle of a Compton camera is as follows: Figure 1a Please refer to Figure 1a After being irradiated by the parallel neutron beam, the radiation source 101 reacts to produce gamma rays, which are captured by the Compton camera. The gamma photons in these gamma rays first interact with the Compton camera's scattering layer 103. If the gamma photons undergo Compton scattering in the scattering layer 103, some of their energy is deposited. The scattered gamma photons then enter the absorption layer 105, where they are completely deposited. This process is called a complete Compton event. During this process, each Compton event corresponds to a Compton cone, whose half-angle is the Compton scattering angle. The angle can be calculated by combining the initial energy of the gamma photons and the deposited energy using a formula. The position of the radiation source 101 is located on the surface of the Compton cone. However, the exact location of the radiation source 101 cannot be determined based on a single Compton cone, or reconstruction cone 107. Numerous Compton events are required to generate multiple reconstruction cones. The intersection of these reconstruction cones is then used to ultimately deduce the specific location of the gamma radiation source.

[0041] Furthermore, Compton camera reconstruction methods can be divided into two categories: one is the traditional back-projection algorithm, and the other is the iterative reconstruction algorithm. The principle of the traditional back-projection reconstruction method is to use the laws of Compton scattering to directly calculate the Compton cone and then obtain the image. Back-projection methods can be divided into two types: the first is the pixel-by-pixel reconstruction method, which divides the space where the radiation source may exist according to a predetermined voxel size, resulting in a three-dimensional imaging space with N×N×M pixels. Alternatively, it can be used to obtain a two-dimensional imaging space with N×N pixels based on the distance in the imaging space. Each voxel in the imaging space is numbered and assigned a weight of 0. Next, each voxel is selected one by one, and the cone surface of each Compton event is calculated to determine whether it passes through the pixel. If so, the weight of the pixel is increased by 1. This process is repeated for all pixels in the space, and the distribution of the radiation source is estimated based on the final weight distribution. When there are enough Compton events used to reconstruct the image, the weight distribution in the imaging space is proportional to the distribution of the radioactive source. The greater the weight in a voxel, the greater the probability that a radioactive source exists in the voxel, and therefore the higher the activity of the radioactive source.

[0042] Although back-projection reconstruction is fast, the spatial resolution of the reconstructed image is poor, making it difficult to meet the requirements of practical applications. Iterative reconstruction algorithms repeatedly correct and converge the data projection and back-projection processes. They are characterized by high spatial resolution of the reconstructed image, but the reconstruction time is much longer than that of back-projection methods. Currently, there are two mainstream iterative algorithms: the stochastic origin ensemble algorithm (SOE) and the maximum likelihood expectation method (MLEM). The stochastic origin ensemble algorithm is also called the origin ensemble algorithm.

[0043] The image reconstruction process of the SOE algorithm is as follows: first, assign a number to the detected event, randomly extract k points from the Compton cone, traverse all events, and obtain the initial probability density distribution of the radiation source. Select event number n, select a sampling point in the reconstruction space, and record the voxel i containing this point. Randomly select a sampling point on the Compton cone of event n, and record the voxel j containing this point. Generate a random number, and determine whether the new random sampling point is accepted based on probability. If accepted, the weight of voxel j is increased by 1. It is understandable that although the SOE algorithm can eliminate more artifacts than the back-projection algorithm, due to its certain limitations, it will cause the signal intensity of some parts of the imaging image to be too high (i.e., concentrated brightness), and the imaging effect needs to be improved. In summary, prompt gamma-single photon emission computed tomography (PG-SPECT) based on a Compton camera is a key technology for achieving online, real-time boron concentration monitoring in BNCT. While the back-projection method based on Compton scattering offers high computational efficiency, the generated images suffer from low spatial resolution and significant artifacts. Iterative algorithms, such as the random source set algorithm (SOE), can improve resolution through iterative probability density optimization, but in practical applications, they are prone to local signal intensity distortion, leading to image contrast imbalance. In other words, the balance between imaging accuracy and computational efficiency in related reconstruction methods has yet to be effectively addressed, resulting in low reliability in the application of these methods in clinical scenarios.

[0044] According to an embodiment of the present application, an embodiment of an image reconstruction method for a Compton camera is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0045] In this embodiment, a method for reconstructing an image of a Compton camera is provided. Figure 1b As shown, the method includes the following steps:

[0046] S110 : Obtain a current weight distribution and a reference weight distribution corresponding to an imaging plane.

[0047] The imaging plane can be a projection plane that visualizes the spatial distribution of a radioactive source in two or three dimensions, representing a slice or level in the Compton camera reconstruction process. For example, upon detection of a Compton event, the space where the radioactive source may be located can be partitioned according to a predetermined voxel size, resulting in multiple N×N two-dimensional imaging planes.

[0048] The current weight distribution can be used to represent the weight of each pixel square within the current imaging plane, representing the relative likelihood of a radioactive source in the reconstructed image. For example, the current weight distribution can be derived using a random source set algorithm. Specifically, based on the principle of Compton scattering, when a gamma-ray photon undergoes a Compton scattering event with the detector of a Compton camera, a Compton cone is generated. The radioactive source is located on the surface of this Compton cone, which is also the reconstruction cone. When this three-dimensional Compton cone is projected onto the two-dimensional imaging plane, it forms a projected ellipse. This ellipse passes through certain pixel squares within the imaging plane, and the weights of these pixel squares are increased accordingly. It is understood that if multiple Compton scattering events occur in space, multiple projected ellipses will be generated. Due to the different positions of the projected ellipses, each pixel square within the imaging plane will also accumulate different weights. These weights represent the likelihood of a radioactive source existing within each pixel square. Furthermore, all detected Compton events are numbered and processed individually. For each event, a random sampling point is selected from the Compton cone, and the pixel square to which it is mapped in the imaging plane is determined. A probabilistic decision is then made as to whether the sampling point is accepted. If accepted, the weight of the pixel square is cumulatively increased. After traversing all events, the weights of the frequently accepted pixel squares in the imaging plane are gradually accumulated to form a probability distribution reflecting the location of the radioactive source. Ultimately, the weight distribution of each pixel square is the current weight distribution, and its value indicates the likelihood of the radioactive source's presence.

[0049] The reference weight distribution can refer to weight data pre-determined based on real radioactive source information. It is used to reflect, under ideal circumstances, the likelihood of a real radioactive source actually existing in each pixel square within the imaging plane. For example, the reference weight distribution can be derived by analyzing and calculating the characteristics of the real radioactive source. Specifically, based on the principle of Compton scattering, if the two-dimensional projected ellipse of each Compton event on the imaging plane passes through certain pixel squares on the imaging plane, the weight of these pixel squares can be increased. After the superposition of multiple events, the accumulated weight values ​​of each pixel square can reflect the spatial probability distribution of the radioactive source, which can be considered the reference weight distribution corresponding to the imaging plane.

[0050] S120 , performing a real radiation source simulation based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane.

[0051] Specifically, the current weight distribution is calculated using a specific iterative algorithm to describe the weight data of the actual radioactive source distribution. However, due to the limitations of the algorithm itself, these data may contain certain errors. Therefore, if the current weight distribution with errors is regarded as the ideal data of the actual radioactive source, and a simulation scenario is constructed based on this weight distribution information to simulate the actual radioactive source, the simulated weight distribution reflecting the relative probability of this simulated radioactive source in each pixel grid on the imaging plane can be obtained.

[0052] Optionally, performing a real radiation source simulation based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane may include generating an inversion circle with a radius of a specified length in the imaging plane based on the current weight distribution, and obtaining the simulated weight value of each pixel square in the imaging plane as the simulated weight distribution corresponding to the imaging plane.

[0053] Specifically, after obtaining the current weight distribution corresponding to the imaging plane, an inversion circle with a specified radius can be generated within the imaging plane based on the current weight distribution. It should be understood that the inversion circle refers to a virtual circular path generated during the Compton camera image reconstruction process based on the relevant characteristics of the Compton scattering event (such as the scattering angle). It approximates the possible projection range of the Compton cone on the imaging plane and can be used to define and simulate the area where a radioactive source may exist. For example, the Compton scattering angle can be first determined based on the initial energy and deposited energy of the gamma photons. This angle is then used to determine the radius of the projected cross-section of the reconstruction cone on the imaging plane, which is used as the specified length to generate the inversion circle.

[0054] Furthermore, the current weight distribution can be used to determine the current weight value of each pixel square on the imaging plane. Inversion circles with random directions and a specified radius passing through each corresponding pixel are generated for each pixel square. The number of inversion circles generated for each pixel square is proportional to the current weight value of that pixel square. Ultimately, the simulated weight value for each pixel square in the inverted imaging plane is obtained. This simulated weight value can be used to reflect the relative probability that each pixel square in the imaging plane may contain a simulated radiation source. The simulated weight values ​​obtained for each pixel square are then combined to form a simulated weight distribution corresponding to the imaging plane.

[0055] It should be noted that because the inversion circle is used to limit the area where the simulated radioactive source may exist, the simulated weight value and simulated weight distribution determined by the inversion circle are different from the current weight distribution that limits the real radioactive source: the simulated weight value and simulated weight distribution are both used to reflect the possibility that a simulated radioactive source may exist in each pixel square in the imaging plane, while the current weight distribution reflects the relative probability of the real radioactive source in each pixel square in the imaging plane.

[0056] S130 , updating the elements in the current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set to perform image reconstruction on the Compton camera.

[0057] The elements in the current source set are used to represent the coordinates of the radiation sources corresponding to Compton events occurring within the Compton camera. Specifically, if each element in the current source set represents the radiation source position corresponding to a Compton event, then the current source set can refer to a collection of radiation source position coordinates corresponding to multiple Compton events, which provides an initial estimate of the radiation source position. For example, in the presence of multiple Compton events, a random point is randomly selected on the Compton cone corresponding to each Compton event, and the projected coordinates of this random point on the imaging plane can be used as an element in the current source set. Each Compton event can correspond to a random point, and the projected coordinates of these random points on the imaging plane constitute the current source set.

[0058] Furthermore, the elements in the current source set are updated according to the simulated weight distribution and the reference weight distribution. First, analysis and processing can be performed based on the simulated weight distribution and the reference weight distribution to locate areas with large weight deviations, so as to guide the replacement and update of the elements in the current source set, and obtain the updated source set, thereby reconstructing the image of the Compton camera.

[0059] Optionally, updating the elements in the current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set may include: determining the current difference data between the simulated weight distribution and the reference weight distribution; updating the elements in the current source set according to the current difference data to obtain an updated source set.

[0060] Specifically, after determining the simulated weight distribution, it is compared with the reference weight distribution corresponding to the imaging plane to determine the current difference data between the two. The current difference data can be used to reflect the degree of deviation between the simulated weight distribution and the reference weight distribution. For example, the current difference data can be calculated by calculating the difference between the simulated weight distribution and the reference weight distribution, thereby serving as an important basis for evaluating and optimizing image reconstruction quality.

[0061] Furthermore, the elements in the current source set are updated based on the current difference data. First, a current element in the current source set can be selected, and another random element can be selected based on the Compton event corresponding to that element. The two elements are then processed separately using the current difference data, and the results are compared and evaluated. If the newly selected random element effectively reduces the deviation, it is used to replace the current element in the current source set. Otherwise, the current element is retained to ensure that the update direction always converges to the high-probability region, ultimately resulting in an updated source set. Similarly, the updated source set can also refer to the new source set obtained by updating the elements in the current source set based on the current difference data during the Compton camera image reconstruction process. Through this update, the coordinates of the radiation sources in the source set can more accurately reflect the actual distribution of the radiation sources, thereby improving the accuracy of image reconstruction. Optionally, the above update process can be iterated, with multiple rounds of sampling and replacement to gradually correct the distribution of the coordinate points in the source set, so that the accumulated weights more closely match the actual radiation source locations, ultimately achieving high-precision imaging results.

[0062] For example, a uniform plane source with 478keV energy can be used as a test source. The geometric arrangement of the plane source is as follows: Figure 1c The plane source can be set 10cm in front of the Compton camera, and the size of the Compton camera can refer to Figure 1d , the Si (silicon) layer can be used as the scattering layer of the Compton camera, and the CZT (cadmium zinc telluride) layer can be used as the absorption layer of the Compton camera. Figure 1e This is the imaging result of the direct back-projection algorithm. It can be seen that there are a lot of artifacts, which are concentrated in the center. Analysis of the results shows a structural contrast of 0.82, a peak noise reduction ratio of 13.64, and a root mean square error of 0.21. Figure 1f The imaging result of the random source set algorithm (SOE) is shown in Figure 1. It can be seen that the central part of the result is more concentrated, and the estimation deviation of the radiation source intensity is larger. The results are also analyzed. The structural contrast of this method is 0.89, the peak noise reduction ratio is 15.20, and the root mean square error is 0.17. The imaging result obtained by the image reconstruction method of the Compton camera of this application is shown in Figure 1. Figure 1gAs shown, it can be seen intuitively that the imaging effect is obvious. While removing artifacts, it also makes the radiation source in the center of the imaging image more concentrated. The imaging results of this method are analyzed, and the structural contrast is 0.92, the peak noise reduction ratio is 18.04, and the root mean square error is 0.12. It can be seen that these three values ​​are better than the direct back projection algorithm and the SOE algorithm. It should be noted that the Stochastic origin ensembles algorithm (SOE) is a Monte Carlo Markov chain (MCMC) method using the Metropolis–Hastings algorithm. In physical applications, the Metropolis–Hastings algorithm is often used to sample complex random distributions. Similarly, the image reconstruction method of the Compton camera of the present application also has similar improvement effects when imaging other types of radiation sources, and can make the center part less concentrated while removing most of the artifacts. For example, the following are reselected: Figure 1h The plane source shown is placed 10 cm in front of the Compton camera. Figure 1i This is the imaging result of the direct back-projection algorithm, with a structural contrast of 0.81, a peak noise reduction ratio of 11.64, and a root mean square error of 0.26. Figure 1j The imaging result of the SOE algorithm has a structural contrast of 0.87, a peak noise reduction ratio of 13.82, and a root mean square error of 0.20. The imaging result obtained by the image reconstruction method of the Compton camera of this application is as follows Figure 1k As shown, its structural contrast is 0.91, the peak noise reduction ratio is 16.88, and the root mean square error is 0.14. It can be seen that the imaging quality of the image reconstruction method of the present application is better than that of the other two algorithms.

[0063] In the above-described embodiment, by obtaining the current weight distribution corresponding to the imaging plane and performing a real radioactive source simulation based on it, a simulated weight distribution is obtained that corresponds to the possible locations of real radioactive sources within the imaged area. Subsequently, the simulated weight distribution and the reference weight distribution are used to locate areas with large weight deviations, and the elements in the current source set are further updated. The elements in the current source set are used to characterize the coordinates of the radioactive sources corresponding to Compton events occurring within the Compton camera. Finally, based on the updated source set, the Compton camera image is reconstructed, so that the reconstructed image more accurately reflects the actual distribution of the radioactive sources, effectively improving the quality and accuracy of the Compton camera image reconstruction, and providing a more reliable basis for real-time monitoring of boron concentration during BNCT treatment, thereby enhancing the safety and effectiveness of BNCT treatment.

[0064] In some embodiments, please refer to the attached Figure 2 , update the elements in the current source set according to the current difference data to obtain the updated source set, including:

[0065] S210: Determine a first probability density of a random element and a second probability density of any element in the current source set.

[0066] Please refer to Figure 1a Based on the imaging principle of the Compton camera, the location of the radiation source 101 is on the Compton cone corresponding to the Compton event, that is, the reconstruction cone 107. However, one Compton event corresponds to only one Compton cone. It is necessary to use multiple Compton events to obtain multiple reconstruction cones, and the specific location of the gamma radiation source can be finally calculated from the intersection of these reconstruction cones.

[0067] The random element is used to represent a random point on the reconstruction cone of the Compton event corresponding to any element. For example, any element can refer to an element already existing in the current source set, representing the possible location of the radiation source considered in the current reconstruction step. The random element can also refer to another randomly selected point on the reconstruction cone of the Compton event corresponding to that element, which can also be used to represent the possible location of the radiation source of that Compton event. Next, the probability of the location of the random element being a radiation source is calculated to obtain a first probability density. The probability of the location of any element being a radiation source is calculated to obtain a second probability density.

[0068] It should be understood that the probability density of an element can refer to the number of voxels corresponding to other elements in the current source set (i.e., a pixel grid in the imaging plane) in the neighborhood of the element. Among them, the neighborhood refers to the area centered on the pixel grid corresponding to a certain element and covering a certain range around the element, which can determine the density of Compton events near the corresponding Compton reconstruction cone. For example, if the element in the current source set corresponds to a pixel grid point in the imaging plane, then the neighborhood can be understood as a circular area with this point as the center, expanding outward with a fixed radius and covering a certain number of surrounding pixel grids.

[0069] S220 . Adjust the first probability density and the second probability density based on the current difference data to obtain a first adjusted density of the random element and a second adjusted density of any element.

[0070] Specifically, the current difference data can first be analyzed to determine the error levels of the random element and any element. Then, based on their respective error levels, calculations are performed to adjust the first probability density corresponding to the random element and the second probability density corresponding to any element, thereby obtaining a first adjusted density reflecting the error significance of the random element and a second adjusted density reflecting the error significance of any element.

[0071] S230 : Update the elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set.

[0072] For example, for a Compton event, the first adjusted density of a random element can be compared with the second adjusted density of any element. If the first adjusted density of the random element is higher, the current element is replaced with the random element to more accurately reflect the actual distribution of the radiation source. Conversely, if the second adjusted density of the current element is higher or equal, the current element is retained. Optionally, the above investigation and replacement process is repeated for all elements in the current source set until all elements in the current source set have been evaluated and updated. This gradually replaces source set elements with elements that closely resemble the actual radiation source distribution, causing the weight distribution to converge toward the actual radiation source location, resulting in an updated source set. Finally, this updated source set is used for image reconstruction, which can significantly improve image reconstruction accuracy.

[0073] In the above implementation, the probability densities of random elements and arbitrary elements are first determined, and then the current difference data is used to adjust and compare the probability densities of each element, thereby optimizing and updating the elements in the source set. This process effectively improves the accuracy of the source set, thereby enhancing the reliability of image reconstruction and ultimately achieving more accurate radiation source location and distribution reconstruction.

[0074] In some embodiments, the current difference data includes the current first-order error distribution and the current second-order error distribution. Figure 3 , adjusting the first probability density and the second probability density based on the current difference data respectively to obtain a first adjusted density of the random element and a second adjusted density of any element, including:

[0075] S310: Use the current first-order error distribution to adjust the second probability density to obtain a second adjusted density of any element.

[0076] Among them, the current first-order error distribution and the current second-order error distribution can both refer to data reflecting the size of the position error of each element in the current source set. For example, the larger the error value, the greater the deviation between the simulated weight at the element position and the reference weight. Furthermore, for any element in the current source set, the second probability density of the any element is adjusted using its corresponding current first-order error distribution, and the second adjusted density of the any element can be obtained. For example, any element in the current source set is denoted as point i, and its probability density, that is, the second probability density, is denoted as In the current difference data, the error coefficient corresponding to any element in the current first-order error distribution is recorded as , then the second adjusted density of any element It can be obtained by the following formula:

[0077]

[0078] in, is the second adjusted density of any element; is the second probability density of any element; is the error coefficient corresponding to any element in the current first-order error distribution.

[0079] S320: Use the current second-order error distribution to adjust the first probability density to obtain a first adjusted density of the random element.

[0080] Similarly, based on any selected element, another random element is selected on the Compton event corresponding to the element, and the first probability density of the random element is adjusted using the current second-order error distribution corresponding to the random element to obtain the first adjusted density of the random element. For example, the selected random element is denoted as point j, and its probability density, that is, the first probability density, is denoted as In the current difference data, the error coefficient corresponding to the random element in the current second-order error distribution is recorded as , then the first adjusted density of the random element It can be obtained by the following formula:

[0081]

[0082] in, is the first adjusted density of the random element; is the first probability density of the random element; is the square of the error coefficient corresponding to the random element in the current second-order error distribution.

[0083] In the above embodiment, the second and first probability densities are adjusted using the current first- and second-order error distributions, respectively, to obtain adjusted densities that are closer to reality. Specifically, the current first-order error distribution is used to adjust the second probability density of any element, reducing the probability density of elements with large errors and increasing the probability density of elements with small errors. The current second-order error distribution is then used to adjust the first probability density of random elements through square amplification, further amplifying the significance of the random element's position. This approach provides a more accurate basis for subsequent source set updates, enabling more precise radioactive source location and distribution reconstruction.

[0084] In some embodiments, updating elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set includes:

[0085] If the first adjusted density is greater than or equal to the second adjusted density, any element is replaced with a random element to obtain an updated source set; if the first adjusted density is less than the second adjusted density, the random element is discarded and any element is retained to obtain an updated source set.

[0086] Specifically, updating the elements in the current source set with the first adjusted density and the second adjusted density can be achieved by comparing the first adjusted density and the second adjusted density. For example, if the first adjusted density of the random element j is Greater than or equal to the second adjusted density of any element i , it means that there are more elements in the current source set that fall within the domain of random element j, that is, the position of random element j is more likely to be the location of the real radiation source, so any element i in the current source set is replaced by random element j to obtain an updated source set that is closer to the real radiation source distribution. Conversely, if the first adjusted density of random element j is Less than the second adjusted density of any element i , it means that in the current source set, more elements fall within the domain of any element i. In other words, the position of any element i is more likely to be the location of the real radioactive source. Therefore, the random element j is discarded and any element i is retained in the current source set to obtain an updated source set that is closer to the real radioactive source distribution.

[0087] Understandably, for the same data, its second-order value is generally smaller than the first-order value. To more accurately capture and reflect the error variations near the position of random elements, first-order difference adjustment is used for any element, while second-order difference adjustment is used for random elements. In other words, only when the result of adjusting the random element for the current second-order error distribution is still greater than or equal to the result of adjusting any element for the current first-order error distribution is the random element considered to have a higher probability of acceptance than any other element and can be replaced, thus screening out higher-quality elements during the source set update process.

[0088] In the above embodiment, the first and second adjusted densities are used to determine which element is most likely to be close to the true radiation source. If the adjusted density of a random element is greater than or equal to any element, that element is replaced to update the source set; otherwise, the original element is retained. This method ensures that the source set retains elements that are more likely to reflect the true radiation source distribution, thereby gradually optimizing the source set and improving the accuracy and reliability of image reconstruction.

[0089] In some embodiments, the simulated weight distribution and the reference weight distribution are obtained through normalization processing, and the current difference data includes the current first-order error distribution and the current second-order error distribution. Figure 4 , the current first-order error distribution and the current second-order error distribution are determined as follows:

[0090] S410 : Subtract the simulated weight distribution from the reference weight distribution to obtain a current error weight distribution as the current first-order error distribution.

[0091] Among them, the current error weight distribution may refer to a set of error coefficients formed by the combination of each pixel in the imaging plane. Exemplarily, the current error weight distribution may be a matrix used to quantify the weight deviation of each pixel. Specifically, after obtaining the reference weight distribution and the simulated weight distribution, they are first normalized to ensure that they can be compared on the same scale, thereby eliminating the influence of dimension. The reference weight distribution is then subtracted from the simulated weight distribution pixel by pixel to obtain the error coefficient of each pixel, thereby forming the current error weight distribution and serving as the current first-order error distribution.

[0092] S420 , performing square calculation on each current error coefficient in the current error weight distribution to obtain a current second-order error distribution.

[0093] The current error coefficient may refer to the weight deviation value of each pixel in the previous error weight distribution, that is, the weight difference value of each pixel. Specifically, each first-order current error coefficient in the current first-order error distribution may be squared to obtain a second-order error coefficient, thereby forming the current second-order error distribution of this imaging plane.

[0094] In this implementation, the normalized reference and simulated weight distributions are used to accurately quantify the weight deviations for each pixel on the imaging plane, forming the current first-order error distribution. Furthermore, by squaring the first-order error coefficients, regions with large errors are highlighted to construct the current second-order error distribution. This process provides critical error information for subsequent optimization of the source set, significantly improving the accuracy and reliability of image reconstruction.

[0095] In some embodiments, please refer to the attached Figure 5 , the specified length is determined as follows:

[0096] S510: Determine an average scattering angle corresponding to a Compton event occurring in the Compton camera, and a distance between a scattering detector of the Compton camera and an imaging plane.

[0097] Specifically, the average scattering angle corresponding to the Compton event occurring in the Compton camera can be calculated based on the Compton scattering formula. For example, it can be calculated using the following formula:

[0098]

[0099] Where θ is the average scattering angle corresponding to the Compton event; is the rest mass of the electron; is the energy of the incident photon; The average scattered energy is calculated by summing and averaging the energy collected by the scattered detector; c is the speed of light.

[0100] Furthermore, the distance between the scattering detector of the Compton camera and the imaging plane may refer to the physical distance between the scattering detector and the imaging plane in the Compton camera. Specifically, the distance may be a predetermined design parameter determined by the geometric layout of the Compton camera. For example, please continue to refer to Figure 1d The 10 cm in the figure can be used as the scattering detector of the Compton camera, that is, the distance between the scattering layer and the imaging plane.

[0101] S520 , determining an average radius of a projection cross section of a reconstructed cone of the Compton event on the imaging plane based on the average scattering angle and the distance, as the designated length.

[0102] Specifically, the projection of the 3D reconstructed cone corresponding to the Compton event on the imaging plane can be considered as a circular area, whose radius is determined by trigonometric calculations based on the scattering angle of the corresponding Compton event and the distance between the scattering detector of the Compton camera and the imaging plane. For example, it can be calculated using the following formula:

[0103]

[0104] Where, is the average radius of the projected cross section of the reconstructed cone of the Compton event on the imaging plane; d is the distance between the scattering detector of the Compton camera and the imaging plane; θ is the average scattering angle corresponding to the Compton event.

[0105] Furthermore, after determining the average radius of the projection section of the reconstructed cone of the Compton event on the imaging plane, it can be used as a specified length to set the range of the inversion circle to simulate the projection range of the Compton cone on the imaging plane, thereby determining the simulation weight distribution corresponding to the imaging plane.

[0106] In the above embodiment, by calculating the average scattering angle of the Compton event and utilizing the known distance between the scattering detector and the imaging plane, the average radius of the projected cross section of the reconstruction cone on the imaging plane is accurately determined, and this average radius is used as the specified length to define the range of the inversion circle, thereby simulating the possible distribution area of ​​the radiation source and performing the image reconstruction process.

[0107] In some embodiments, the reference weight distribution is used to describe the first weight value of each pixel square in the imaging plane obtained by the analytical reconstruction algorithm, and the current weight distribution is the second weight value of each pixel square in the imaging plane in the current round of the iterative reconstruction algorithm.

[0108] If the current round is the first round, the current weight distribution is obtained through the random source set algorithm; if the current round is not the first round, the current weight distribution is obtained by weight matching the updated source set obtained in the previous round of the current round with the pixels in the imaging plane.

[0109] The analytical reconstruction algorithm may refer to an algorithm that uses direct mathematical operations to calculate an image in one go based on an imaging model and measurement data. For example, the reference weight distribution may refer to the first weight values ​​of each pixel square in the imaging plane obtained by the direct back-projection algorithm within the analytical reconstruction algorithm. The direct back-projection algorithm is a method that directly projects the Compton cone onto the imaging plane and assigns weight values ​​to the covered pixels.

[0110] An iterative reconstruction algorithm can be one that progressively approximates the true image through a repetitive optimization process. Starting from an initial estimate, it performs multiple rounds of iterative corrections, with each iteration bringing the image closer to the measured data and conforming to physical laws. For example, an iterative reconstruction algorithm can include a random source set algorithm, which suppresses artifacts by dynamically adjusting the source set sampling points and combining error feedback.

[0111] It can be understood that if the current round is the first, the goal of this round is to generate initial weight distribution data. First, based on the random source set algorithm, all detected Compton events are numbered and processed one by one. For each event's corresponding Compton cone, a sampling point is randomly selected, and the pixel square to which the point is mapped in the imaging plane is determined. A probabilistic decision is then made as to whether the sampling point is accepted. If accepted, the weight value of the pixel square is cumulatively increased. After traversing all events, the weights of the frequently accepted pixel squares in the imaging plane are gradually accumulated to form initial weight distribution data reflecting the location of the radiation source, which serves as the current weight distribution for the current round (first round). Next, if the current round is the first, this initial weight distribution data is used to generate an inversion circle and further obtain a simulated weight distribution corresponding to the imaging plane. The current difference between the simulated weight distribution and the reference weight distribution corresponding to the imaging plane is then calculated, and the elements in the current source set are updated based on this difference data to obtain the updated source set.

[0112] Furthermore, based on the iterative reconstruction algorithm, the updated source set is used to map each element in the source set to each pixel on the imaging plane, the number of coverages is counted and the updated weight distribution is regenerated as the current weight distribution of the current round (second round).

[0113] In this implementation, the image quality of the Compton camera is significantly improved by combining the rapidity of the analytical reconstruction algorithm with the iterative optimization capabilities of the iterative reconstruction algorithm. In the first round, a random source set is used to generate an initial weight distribution. Subsequently, the source set sampling strategy is dynamically adjusted using error coefficients (first-order and second-order errors), prioritizing correction of high-deviation areas, thereby gradually guiding the algorithm to focus on the true radiation source location. Through multiple rounds of iteration, the sampling points in the source set gradually converge to high-probability areas. The resulting weight distribution effectively suppresses artifacts and significantly improves image quality.

[0114] The present disclosure also provides a method for reconstructing an image from a Compton camera, the method comprising the following steps:

[0115] S602 : Obtain a first weight value of each pixel grid in the imaging plane through an analytical reconstruction algorithm as a reference weight distribution.

[0116] S604: Obtain a second weight value for each pixel grid in the imaging plane using a random source set algorithm as the current weight distribution. The random source set algorithm includes randomly selecting sampling points on the cone of the Compton event.

[0117] S606: Determine an average scattering angle corresponding to a Compton event occurring in the Compton camera, and a distance between a scattering detector of the Compton camera and an imaging plane.

[0118] S608 . Determine an average radius of a projection cross section of a reconstructed cone of the Compton event on the imaging plane based on the average scattering angle and the distance, as the designated length.

[0119] S610 , the current weight distribution generates an inversion circle with a specified radius in the imaging plane, and obtains a simulation weight value of each pixel grid in the imaging plane as the simulation weight distribution corresponding to the imaging plane.

[0120] S612 : Subtract the normalized simulation weight distribution from the normalized reference weight distribution to obtain a current error weight distribution as the current first-order error distribution.

[0121] S614 , performing square calculation on each current error coefficient in the current error weight distribution to obtain a current second-order error distribution.

[0122] S616: Select a random element and any element in the current source set. The element in the current source set is used to represent the coordinates of the radiation source corresponding to a Compton event occurring within the Compton camera. The random element is used to represent a random point on the reconstruction cone of the Compton event corresponding to the element.

[0123] S618: Determine a first probability density of the random element and a second probability density of any element in the current source set.

[0124] S620: Use the current first-order error distribution to adjust the second probability density to obtain a second adjusted density of any element.

[0125] S622: Use the current second-order error distribution to adjust the first probability density to obtain a first adjusted density of the random element.

[0126] S624: If the first adjusted density is greater than or equal to the second adjusted density, replace any element with a random element; otherwise, if the first adjusted density is less than the second adjusted density, discard the random element and retain any element to obtain an updated source set for the current round.

[0127] S626 , using the updated source set obtained in step S624 as the current source set in step S616 , and repeating steps S616 to S624 to obtain an iterative updated source set.

[0128] S628 . Perform weight matching on the pixels in the imaging plane using the iteratively updated source set to obtain the iterative weight distribution of the current round.

[0129] S630: Use the iterative weight distribution of the current round as the current weight distribution of the next round in step S604, and repeat steps S604 to S628 to obtain a reconstructed weight distribution.

[0130] S632. After obtaining a reconstruction weight distribution, start the first iteration of a new cycle, reselect sampling points based on the random source set algorithm to re-determine the current weight distribution of the imaging plane, and repeat steps S604 to S628 to obtain multiple reconstruction weight distributions.

[0131] S634: Perform weighted averaging processing on the multiple reconstruction weight distributions to obtain a target weight distribution.

[0132] S636: Reconstruct the image using the target weight distribution to obtain a high-quality reconstructed image.

[0133] It should be understood that, although the various steps in the above flowchart are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be performed in other orders. Moreover, at least a portion of the steps in the above flowchart may include multiple steps or multiple stages, and these steps or stages are not necessarily performed at the same time, but can be performed at different times. The execution order of these steps or stages is not necessarily to be performed in sequence, but can be performed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0134] The embodiment of this specification also provides an image reconstruction device 600 of a Compton camera, such as Figure 6 As shown, it includes: a weight distribution acquisition module 610, a simulation weight determination module 620 and a current source set update module 630. Among them:

[0135] The weight distribution acquisition module 610 is used to acquire the current weight distribution and the reference weight distribution corresponding to the imaging plane.

[0136] The simulation weight determination module 620 is used to perform a real radiation source simulation based on the current weight distribution to obtain a simulation weight distribution corresponding to the imaging plane.

[0137] The current source set update module 630 is used to update the elements in the current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set for image reconstruction of the Compton camera; wherein the elements in the current source set are used to represent the coordinates of the radiation source corresponding to the Compton event occurring in the Compton camera.

[0138] In some embodiments, the current source set update module 630 is also used to update the elements in the current source set according to the simulated weight distribution and the reference weight distribution to obtain an updated source set, including: determining the current difference data between the simulated weight distribution and the reference weight distribution; updating the elements in the current source set according to the current difference data to obtain an updated source set.

[0139] In some embodiments, the simulation weight determination module 620 is also used to simulate a real radiation source based on the current weight distribution to obtain a simulation weight distribution corresponding to the imaging plane, including: generating an inversion circle with a radius of a specified length in the imaging plane based on the current weight distribution, and obtaining the simulation weight value of each pixel square in the imaging plane as the simulation weight distribution corresponding to the imaging plane.

[0140] In some embodiments, the current source set update module 630 is also used to update the elements in the current source set according to the current difference data to obtain an updated source set, including determining a first probability density of the random element and a second probability density of any element in the current source set; wherein the random element is used to characterize a random point on the reconstructed cone of the Compton event corresponding to any element; based on the current difference data, the first probability density and the second probability density are adjusted respectively to obtain a first adjusted density of the random element and a second adjusted density of any element; the elements in the current source set are updated according to the first adjusted density and the second adjusted density to obtain an updated source set.

[0141] In some embodiments, the current difference data includes a current first-order error distribution and a current second-order error distribution. The current source set update module 630 is further configured to adjust the first probability density and the second probability density based on the current difference data to obtain a first adjusted density of the random element and a second adjusted density of any element, including: adjusting the second probability density using the current first-order error distribution to obtain the second adjusted density of any element; and adjusting the first probability density using the current second-order error distribution to obtain the first adjusted density of the random element.

[0142] In some embodiments, the current source set update module 630 is further used to update the elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set, including: if the first adjusted density is greater than or equal to the second adjusted density, replacing any element with a random element to obtain an updated source set; if the first adjusted density is less than the second adjusted density, discarding the random element and retaining any element to obtain an updated source set.

[0143] In some embodiments, the simulated weight distribution and the reference weight distribution are each obtained through normalization, and the current difference data includes a current first-order error distribution and a current second-order error distribution. Difference data determination module 620 is further configured to determine the current first-order error distribution and the current second-order error distribution by: subtracting the simulated weight distribution from the reference weight distribution to obtain a current error weight distribution as the current first-order error distribution; and squaring each current error coefficient in the current error weight distribution to obtain a current second-order error distribution.

[0144] In some embodiments, the weight distribution acquisition module 610 is further used to determine the specified length by: determining the average scattering angle corresponding to the Compton event occurring in the Compton camera, and the distance between the scattering detector of the Compton camera and the imaging plane; based on the average scattering angle and the distance, determining the average radius of the projected cross-section of the reconstructed cone of the Compton event on the imaging plane as the specified length.

[0145] In some embodiments, the reference weight distribution is used to describe the first weight values ​​of each pixel square in the imaging plane obtained by the analytical reconstruction algorithm, and the current weight distribution is the second weight value of each pixel square in the imaging plane in the current round of the iterative reconstruction algorithm. The weight distribution acquisition module 610 is further configured to determine whether, if the current round is the first round, the current weight distribution is obtained by using a random source set algorithm; if the current round is not the first round, the current weight distribution is obtained by weight matching the updated source set obtained in the round before the current round with the pixels in the imaging plane.

[0146] The specific definition of a Compton camera image reconstruction device can be found in the definition of a Compton camera image reconstruction method described above and will not be further elaborated here. Each module in the Compton camera image reconstruction device described above can be implemented in whole or in part via software, hardware, or a combination thereof. Each of these modules can be embedded in or independent of a processor in a computer device in hardware form, or stored in a computer device memory in software form, allowing the processor to call and execute the corresponding operations of each module.

[0147] In this embodiment, an image reconstruction device of a Compton camera is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0148] The embodiment of the present application further provides a computer device, which may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 7 As shown. The computer device includes a processor, memory, a communication interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal via wired or wireless communication. The wireless communication can be achieved via Wi-Fi, a mobile cellular network, NFC (near-field communication), or other technologies. When executed by the processor, the computer program implements a Compton camera image reconstruction method. The display screen of the computer device can be a liquid crystal display or an electronic ink display. The input device of the computer device can be a touch layer covering the display screen, or keys, a trackball, or a touchpad provided on the computer device housing, or an external keyboard, touchpad, or mouse.

[0149] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0150] The embodiments of the present application also provide a computer-readable storage medium. The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0151] An embodiment of the present application provides a computer program product, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform a method according to any embodiment of the present application.

[0152] The Compton camera image reconstruction method, apparatus, computer device, and storage medium described in the above embodiments can be implemented by a computer chip or physical device, or by a product with certain functions. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. Throughout this specification, references to the terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" indicate that the specific features, structures, materials, or characteristics described in conjunction with such embodiment or example are included in at least one embodiment or example of the present application. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed to indicate or imply relative importance or to implicitly specify the quantity of the technical features indicated. Thus, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined. It should also be noted that the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, commodity or equipment comprising a series of elements includes not only those elements, but also includes other elements that are not explicitly listed, or also includes elements inherent to such process, method, commodity or equipment. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the presence of other identical elements in the process, method, commodity or equipment comprising the elements.

[0153] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. Since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment. The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of the claims of the present application. Although the embodiments of the present application are described in conjunction with the drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the present application, and such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A Compton camera image reconstruction method, characterized in that: The method comprises: Obtaining the current weight distribution and the reference weight distribution corresponding to the imaging plane; Performing a real radiation source simulation based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane; Determining current difference data between the simulated weight distribution and the reference weight distribution, and determining a first probability density of a random element and a second probability density of any element in the current source set; wherein the current difference data includes a current first-order error distribution and a current second-order error distribution; and the random element is used to characterize a random point on a reconstructed cone of a Compton event corresponding to any element; Adjusting the first probability density using the current second-order error distribution to obtain a first adjusted density of the random element; adjusting the second probability density using the current first-order error distribution to obtain a second adjusted density of the any element; The elements in the current source set are updated according to the first adjusted density and the second adjusted density to obtain an updated source set for performing image reconstruction on the Compton camera; wherein the elements in the current source set are used to represent the coordinates of the radiation source corresponding to the Compton event occurring in the Compton camera.

2. The method according to claim 1, characterized in that The performing of a real radiation source simulation based on the current weight distribution to obtain a simulated weight distribution corresponding to the imaging plane includes: An inversion circle with a specified radius is generated in the imaging plane based on the current weight distribution, and a simulation weight value of each pixel grid in the imaging plane is obtained as the simulation weight distribution corresponding to the imaging plane.

3. The method according to claim 1, characterized in that Updating the elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set, including: If the first adjusted density is greater than or equal to the second adjusted density, any one of the elements is replaced by the random element to obtain the updated source set.

4. The method according to claim 1, wherein Updating the elements in the current source set according to the first adjusted density and the second adjusted density to obtain an updated source set, including: If the first adjusted density is less than the second adjusted density, the random element is discarded and any one element is retained to obtain the updated source set.

5. The method according to claim 1, characterized in that The simulated weight distribution and the reference weight distribution are obtained by normalization processing respectively; the current first-order error distribution is determined by the following method: The simulation weight distribution is subtracted from the reference weight distribution to obtain a current error weight distribution as the current first-order error distribution.

6. The method according to claim 5, characterized in that The current second-order error distribution is determined by: Each current error coefficient in the current error weight distribution is squared to obtain the current second-order error distribution.

7. The method according to claim 2, characterized in that The specified length is determined as follows: determining an average scattering angle corresponding to Compton events occurring in the Compton camera and a distance between a scattering detector of the Compton camera and the imaging plane; An average radius of a projection cross section of a reconstructed cone of the Compton event on the imaging plane is determined based on the average scattering angle and the distance as the designated length.

8. The method according to any one of claims 1 to 7, characterized in that The reference weight distribution is used to describe the first weight value of each pixel square in the imaging plane obtained by the analytical reconstruction algorithm, and the current weight distribution is the second weight value of each pixel square in the imaging plane in the current round of the iterative reconstruction algorithm; If the current round is the first round, the current weight distribution is obtained by a random source set algorithm; If the current round is not the first round, the current weight distribution is obtained by weight matching the updated source set obtained in the round before the current round with the pixels in the imaging plane.

9. An image reconstruction device for a Compton camera, characterized in that: The device comprises: A weight distribution acquisition module is used to obtain the current weight distribution and reference weight distribution corresponding to the imaging plane; a simulation weight determination module, configured to perform a real radiation source simulation based on the current weight distribution to obtain a simulation weight distribution corresponding to the imaging plane; A current source set update module is used to determine the current difference data between the simulated weight distribution and the reference weight distribution, and to determine the first probability density of the random element and the second probability density of any element in the current source set; wherein the current difference data includes the current first-order error distribution and the current second-order error distribution; the random element is used to characterize the random point on the reconstruction cone of the Compton event corresponding to any element; the first probability density is adjusted using the current second-order error distribution to obtain the first adjusted density of the random element; the second probability density is adjusted using the current first-order error distribution to obtain the second adjusted density of any element; the elements in the current source set are updated according to the first adjusted density and the second adjusted density to obtain an updated source set, so as to reconstruct the image of the Compton camera; wherein the elements in the current source set are used to characterize the coordinates of the radiation source corresponding to the Compton event occurring in the Compton camera.

10. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method according to any one of claims 1 to 8 by executing the computer instructions.

11. 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 computer to execute the method according to any one of claims 1 to 8.

Images