Acquisition method for detecting real PSF function by using single-source SPECT

By performing actual measurements on the SPECT detector to obtain the real PSF function, the problem of unreality of simulated PSF function and large computing resource utilization is solved, the accuracy and efficiency of image reconstruction are improved, and the real reflection of the actual environment of the instrument is realized.

CN119001815APending Publication Date: 2024-11-22SHANDONG MADIC TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411103137.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-12
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

In the existing SPECT technology, the PSF function is mainly obtained through simulation methods, which fails to truly reflect the actual environment of the instrument, resulting in poor image reconstruction effect, large computing resources occupies and systematic errors.

Method used

By conducting actual measurements on the gate format SPECT detector equipped with collimator, the square source is configured to traverse each detector grid, record radiation information and its acquisition time, split and merge PSF change curve data, and obtain real PSF accumulation Function, used for image reconstruction.

Benefits of technology

It realizes the acquisition of real PSF functions, improves the accuracy and efficiency of image reconstruction, avoids waste of computing resources, can objectively reflect the actual situation of the instrument, and reduces system errors.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119001815A_ABST
    Figure CN119001815A_ABST
Patent Text Reader

Abstract

A real PSF function acquisition method comprises the following steps: 1) configuring a square source of which the sectional area is not smaller than a detector crystal grid for a grid type SPECT of a registration collimator; 2) during detection, enabling the square source to be located at a certain layer in Z layers in the Z-axis direction, setting the translation speed of the square source, keeping the translation speed to move in an X-Y plane, traversing each detector grid to emit radiation, and recording radiation information and obtaining time by the detector; repeating the operation for other layers (Z-1); (3) aiming at a certain layer in Z layers in the Z-axis direction, aiming at the information obtained in the step (2), segmenting radiation information data, and obtaining a PSF change curve in the X-axis direction and the Y-axis direction when the source is located above the grid of the current detector; performing the same operation on other X * Y-1 detector grids in the current layer; (4) repeating the step (3) for other Z-1 layers; and 5) combining all PSF change curve data to obtain a real PSF accumulation function of the detector.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the technical field of PSF function acquisition in SPECT technology, specifically a method for obtaining the real PSF function using a single-source SPECT probe. Background Art

[0002] SPECT (Single-Photon Emission Computed Tomography) is one of the two ECT (Emission Computed Tomography) techniques in nuclear medicine imaging. It uses radiopharmaceuticals that emit only a single photon with each decay to create tomographic images, revealing the distribution of the drug within the organism and reflecting its functional, metabolic, and physiological state. Compared to PET, SPECT has both advantages and disadvantages; overall, it is one of the more important methods for detecting nuclear medicine imaging and is widely used in clinical settings.

[0003] Besides its clinical applications, SPECT is now widely used in small animal drug research. Due to the increased safety margin, small animals can withstand more radiation / exposure than human medical studies, and SPECT has fewer limitations in small animal drug research. Therefore, small animal SPECT is widely used in clinical research for novel drugs and is frequently used for the detection and research of laboratory animals.

[0004] To generate images, SPECT technology must acquire them through an iterative reconstruction algorithm. Two types of data are essential: the projected image used in the iterative reconstruction algorithm and the system point spread function (PSF) required by the algorithm. (See reference 1, Parallel Monte Carlo Simulation of MicroSPECT Imaging System Based on GATE Program, Journal of Tsinghua University (Natural Science Edition), 2007, Vol. 47, No. S1). Intuitively, the PSF function can be understood as... Figure 4 The trend of PSF on the X-axis as a function of the x-coordinate of the point source, and Figure 5 The PSF function is the trend of the PSF on the Y-axis as the y-coordinate of the point source changes, combined with the changes in distance along the z-axis (for example, the intensity is greater for power sources closer to each other, and less for those farther away).

[0005] However, current PSF functions are all obtained through simulation. For example, for a company's ABC type spectral, the PSF function of this type of machine is obtained by simulation calculation in a computing center according to the settings. As mentioned in Reference 1, "In the GATE program, by uniformly placing 64×64×64=262144 point sources in the field of view, tracking a large number of X-ray photons emitted from each point source and recording the X-ray photon counts received on each detection unit, the PSF of the MicroSPECT system can be simulated." "Because the simulation speed of the GATE program is relatively slow, parallel computing is required on the Tsinghua High-Performance Cluster (THPCC) to improve the computing speed. Since the GATE program itself does not support parallel computing, parallelization of the simulation calculation is achieved by running the GATE program simultaneously with multiple tasks and setting different initial random number seeds for each task."

[0006] Obtaining the PSF function through simulation has several problems. First, it doesn't rely on actual measurements or consider the instrument's real-world environment. Therefore, compared to experimentally measured PSF functions, its image reconstruction performance is inferior, and it has an upper limit due to systematic errors. Some practical factors cannot be considered or incorporated into the simulation's generation. Second, it relies heavily on computing resources. Such a large computational load typically requires large computing centers like Tsinghua University's High Performance Cluster (THPCC). In contrast, experimentally measured PSF functions require recording data in memory during measurement, and then processing and transforming the data afterward to obtain the instrument's PSF function, without requiring a large computational load. Thirdly, and most importantly, the existing PSF function is a predictive function and not truly accurate. This can lead to problems with the normalization of SPECT detectors, causing discrepancies between the actual photon emission and detection positions. For example, suppose the signal strength is highest in row 9, but in actual measurements, the signal attenuation towards rows 8, 7, 10, and 11 is very steep (the changes are more abrupt). This is difficult to reflect in analog signals. Therefore, applying the analog PSF to an actual machine may result in some image areas being too dark during reconstruction, making it difficult to identify or display effective information. Furthermore, the measured PSF function can objectively reflect the influence of various factors encountered by the instrument in its local environment. For example, the local influences of radiation and electrical systems, and even the actual effects of air pressure, temperature, and humidity, cannot be reflected in the analog PSF function.

[0007] In summary, the purpose of this application is to completely differ from the existing technology in the acquisition method of generating simulated PSF functions. It conducts actual measurements on specific Spectrum instruments and ensures that subsequent Spectrum detections are based on the actual acquired PSF function to generate reconstructed images of the Spectrum detection results. Summary of the Invention

[0008] The purpose of this invention is to provide a method that addresses a series of problems in existing spectral detection methods, such as reliance on simulated PSF functions. These simulated PSF function generation does not consider the actual instrument environment, consumes significant computational resources, and suffers from poor image reconstruction results due to discrepancies between the simulated function and the actual situation. This method is original and not inspired by any prior art. Through careful design, the method of this application offers numerous advantages. Based on actual instrument measurements, it overcomes the problems of unrealistic and unrealistic simulated PSF functions, effectively ensuring the image reconstruction performance of current instruments.

[0009] To achieve the above objectives, the present invention provides the following technical solution: a method for obtaining a real PSF function, characterized in that: (1) for a grid-type SPECT detector equipped with a collimator, a square source with a cross-sectional area not less than the detector crystal grid is configured; the grid-type SPECT detector has X*Y grids.

[0010] (2) Keep the SPECT detector in the detection state, place the source in a certain layer within Z layers in the Z-axis direction, set its translation speed, keep the translation speed in the XY plane, move the source and emit radiation by traversing at least once the path of each detector grid, and the detector records all radiation information and its acquisition time.

[0011] For the other Z-1 layers, change the source to the corresponding height and repeat the above operation to obtain radiation information and the acquisition time.

[0012] (3) For a certain layer within the Z layers in the Z-axis direction, based on the radiation information obtained in step (2) and its acquisition time, the radiation information data is segmented according to the motion path of the source and the time node when it moves into or out of a detector grid. For each grid, the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction when the source is located above the current detector grid are obtained.

[0013] For the other X*Y-1 detector grids within the same layer, perform the same operation to obtain their corresponding PSF variation curves in the X-axis and Y-axis directions.

[0014] 4) For the other Z-1 layers, repeat the same operation in step (3) to obtain the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction for each detector grid in each layer.

[0015] 5) Combine all PSF change curve data from steps (3) and (4) to obtain the true PSF accumulation function of the corresponding SPECT detector.

[0016] A SPECT imaging method is characterized by: using the real PSF function acquisition method as described above for current SPECT detection equipment, combined with the projection data obtained from a single detection, reconstructing the original radioactivity distribution, and then obtaining an image that can be displayed.

[0017] This application also discloses another method for obtaining the real PSF function, characterized in that: (1) for a grid-type SPECT detector equipped with a collimator, a square source with a cross-sectional area not less than the detector crystal grid is configured; the grid-type SPECT detector has X*Y grids.

[0018] (2) Keep the SPECT detector in the detection state, place the source in a certain layer within Z layers in the Z-axis direction, set its translation speed, keep the translation speed in the XY plane, move the source and emit radiation by traversing at least once the path of each detector grid, and the detector records all radiation information and its acquisition time.

[0019] For the other Z-1 layers, change the source to the corresponding height and repeat the above operation to obtain radiation information and the acquisition time.

[0020] (3) For a certain layer within the Z layers in the Z-axis direction, based on the radiation information obtained in step (2) and its acquisition time, the radiation information data is segmented according to the motion path of the source and the time node when it moves into or out of a detector grid. For each grid, the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction when the source is located above the current detector grid are obtained.

[0021] For the other X*Y-1 detector grids within the same layer, perform the same operation to obtain their corresponding PSF variation curves in the X-axis and Y-axis directions.

[0022] (4) For the other Z-1 layers, repeat the same operation in step (3) to obtain the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction for each detector grid in each layer.

[0023] 5) Collect all PSF change curve data from steps (3) and (4), and replace one or more unreasonable PSF change curve data with simulated PSF change curve data.

[0024] 6) Merge all PSF change curve data collected and replaced in step (5) to obtain the true PSF accumulation function of the corresponding SPECT detector.

[0025] Furthermore, the simulated PSF change curve data is selected from the following: the expected value of the PSF change curve data collected at this location, the estimated value based on data from 4-16 adjacent locations, and the measured value of the PSF change curve data collected at this location in a previous instance.

[0026] The rules for judging unreasonable PSF change curve data are as follows: For a certain PSF change curve, based on the measured data of the PSF change curves at its adjacent 4-16 positions, its mathematical expectation value is obtained. If the PSF change curve data collected for that position is compared with the mathematical expectation value and is generally lower by A% or higher by A% or more, it is considered unreasonable, where A∈[1,10].

[0027] A SPECT imaging method is characterized by: using the real PSF function acquisition method as described above for current SPECT detection equipment, combined with the projection data obtained from a single detection, reconstructing the original radioactivity distribution, and then obtaining an image that can be displayed.

[0028] Compared to existing technologies, this application has several non-obvious advantages: First, the acquisition of the PSF real function in this application relies entirely on actual measurements, resulting in a true PSF function. The signal strength received from the corresponding source is entirely realistic. If the actual signal strength is stronger or weaker than predicted, this will also be reflected in the real function, allowing the reconstructed image to better reflect the real situation. Second, compared to the computationally intensive simulation of the PSF function, which relies heavily on large computing centers like Tsinghua High-Performance Cluster (THPCC), the real PSF function only requires actual data acquisition without a large computational requirement. This avoids resource waste. Large computing resources, whether purchased or rented, are extremely expensive, significantly increasing equipment costs. The actual increase in this application is only the equipment installation and debugging time. Generally, Spectrum equipment debugging takes half a month; adding 7-10 days to that, plus the data sorting and review time for one staff member, is sufficient to complete the processing of the real PSF function, saving resources. Third, a real PSF can capture data from the actual conditions of the equipment. For example, there may be discrepancies between the actual photon emission position and the detection position of a real SPECT device. For instance, suppose the signal strength is highest in line 9, but in actual measurements, the signal attenuation towards lines 8, 7, 10, and 11 is very steep (the changes are more sudden). This is difficult to reflect in analog signals, but real PSF does not have this problem; this trend is fully reflected in real PSF. This avoids the distortion problem after conversion from analog PSF. Real PSF is actually acquired and completely corresponds to the actual conditions of the equipment (the source, scintillation crystal array, and sometimes the properties of the crystal blocks in the crystal array differ). Fourth, the measured PSF function can also actually receive the influence of various factors encountered by the instrument on-site, all of which are objectively reflected in the measured function. For example, the various effects of local radiation and electrical conditions on the instrument, and even the actual effects that air pressure, temperature, and humidity may cause, are all things that would normally take half a month to debug a typical SPECT device. This can be achieved by extending the measurement by 7-10 days. This ability to eliminate potential systematic errors better ensures the quality of image reconstruction, which is something that analog PSF cannot provide. Attached Figure Description

[0029] Figure 1 This diagram illustrates the fixed and mobile operation methods of the source.

[0030] Figure 2 This is a schematic diagram of the source moving scan method.

[0031] Figure 3 The flowchart for acquiring data by moving the source scan.

[0032] Figure 4This is a graph showing the changes in the acquired signals of a certain crystal XiYi at layers 0 and 20.

[0033] Figure 5 This is a schematic diagram showing the positional segmentation of the single-layer scanning plane of the Fangyuan source.

[0034] Reference numerals: 1. Source, 2. 3D module, 21. X-axis module, 22. Y-axis module, 23. Z-axis module, 3. Collimator, 4. FOV range, 51. Source start point, 52. Source end point, 53. Collimator plane, 54. Scanning plane, 55. X42 detection row, 56. Y22 detection column, 57. X42Y22 crystal position. DETAILED DESCRIPTION

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Example 1

[0037] The overall objective of this invention is to determine the true PSF function and normalize the detector using this true PSF function. Specifically, it aims to obtain the true PSF function corresponding to each voxel point (voxel size is resolution-dependent) within the detection space. The signal is obtained, for example, by using a moving radiation source with an area consistent with the detector's surface area. The moving radiation source is, for example, cube / cuboidal in shape and can be any conventionally available radiation source. The detection method involves placing the point source above the detector, adjusting the point source position, detecting the detector's count when the point source is located at each voxel point, and fitting the PSF function of the corresponding voxel point using the count data.

[0038] A method for obtaining a real PSF function, characterized in that: (1) for a grid-type SPECT detector equipped with a collimator, a square source with a cross-sectional area not less than the detector crystal grid is configured; the grid-type SPECT detector has X*Y grids. Here, X and Y generally refer to positive integers, and depending on the actual situation, such as values ​​between 15 and 100, they can be set equally, such as 16*16, 25*25, 30*30, 40*40, 60*60, 80*80 grid arrays, or they can be set asymmetrically, such as a 60*80 grid array.

[0039] (2) Keep the SPECT detector in the detection state, place the source in a certain layer within Z layers in the Z-axis direction, set its translation speed, keep the translation speed in the XY plane, move the source and emit radiation by traversing at least once the path of each detector grid, and the detector records all radiation information and its acquisition time.

[0040] Here, the most common way to start a layer is from the bottom or top, such as from layer 1, layer 0, or layer 50 or layer 100. However, this application does not limit the initial layer selection and any layer can be chosen to start.

[0041] This refers to "moving the source and emitting radiation along a path that traverses each detector grid at least once." This path can be chosen based on the specific circumstances. For example, it could involve continuously circling from one corner to the center grid, circling counter-clockwise or clockwise from the center grid to the outermost corner grid, or using an interlaced turning motion, such as moving from grid A1 to A22, then to B22, then back to B1, and then from the adjacent C1 to C22, and so on until the entire process is complete. Due to accidental repetitions or limitations of the moving device, one or more grids may be traversed more than once. In this case, only the data from the first traversal needs to be collected or retained; this can be chosen based on the specific data.

[0042] For the other Z-1 layers, change the source to the corresponding height and repeat the above operation to obtain radiation information and the acquisition time. The detection of all Z layers can be sequential, such as from 1 to Z, or from Z to 1, or the order can be randomly selected, depending on the operator's preference or actual needs.

[0043] (3) For a certain layer within Z layers along the Z-axis, based on the radiation information obtained in step (2) and its acquisition time, the radiation information data is segmented according to the motion path of the source and the time node when it moves into or out of a detector grid. For each grid, the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction are obtained when the source is above the current detector grid. Here, the PSF change curves in the X-axis direction and the PSF change curves in the Y-axis direction are both planar change information, which is a common data recording format. Of course, according to the actual situation, such as the current database supporting three-dimensional data, the PSF change curves in the XY-axis direction can be recorded directly. For example, the illustration is a three-dimensional stereoscopic image. Each node (grid) generates a bar-shaped peak, a column-shaped peak, etc., based on the cumulative height. If it is the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction, the actual illustration is, for example, the shape of a continuous mountain range, with the highest peak in the middle and the height decreasing outwards (this is a general trend, and the measured curve is definitely significantly different from the simulated curve. Simply put, it is very likely that this curve is not symmetrical).

[0044] The same operation is performed on the other X*Y-1 detector grids within the same layer to obtain their corresponding PSF variation curves in the X and Y axes. This calculation method is either exactly the same or similar, but for the grids in the first or second row at the edge, a slightly different calculation method may be used.

[0045] 4) For the other Z-1 layers, repeat step (3) to obtain the PSF variation curves in the X-axis direction and the PSF variation curves in the Y-axis direction for each detector grid in each layer. The recording method for this operation is the same, and the data is recorded in the same way.

[0046] 5) Combine all PSF change curve data from steps (3) and (4) to obtain the true PSF accumulation function of the corresponding SPECT detector. This data calculation can be achieved by first establishing a mapping of function relationships for a predetermined format, then directly filling in the received data, or by filling the data into predetermined data slots according to certain specific formats, such as data corresponding to a certain source or a certain grid, and then automatically performing calculations based on the mapping relationships set in the database.

[0047] A SPECT imaging method is characterized by: for current SPECT detection equipment, utilizing the previously described real PSF function acquisition method, combined with projection data obtained from a single detection, to reconstruct the original radioactivity distribution, and thus obtain a displayable image. The reconstructed original radioactivity distribution here is entirely consistent with reality, unlike simulations which are based on the equipment operating under completely ideal conditions, generally exhibiting significant differences. Common problems include overall overexposure, local overexposure, overall underexposure, and local underexposure. The real PSF function of this application can, in most cases, avoid image reconstruction distortion caused by overly idealized simulated PSF functions, making observation difficult and failing to accurately reflect the real situation and true image.

[0048] Example 2

[0049] 5) Collect all PSF change curve data from steps (3) and (4), and replace one or more unreasonable PSF change curve data with simulated PSF change curve data.

[0050] 6) Merge all PSF change curve data collected and replaced in step (5) to obtain the true PSF accumulation function of the corresponding SPECT detector.

[0051] Furthermore, the simulated PSF change curve data is selected from the following: the expected value of the PSF change curve data collected at this location, the estimated value based on data from 4-16 adjacent locations, and the measured value of the PSF change curve data collected at this location in a previous instance.

[0052] The rules for judging unreasonable PSF change curve data are as follows: For a certain PSF change curve, based on the measured data of the PSF change curves at its adjacent 4-16 positions, its mathematical expectation value is obtained. If the PSF change curve data collected for that position is compared with the mathematical expectation value and is generally lower by A% or higher by A% or more, it is considered unreasonable, where A∈[1,10].

[0053] This embodiment is based on the consideration that it uses actual measured values ​​instead of a simulated PSF function. However, in actual measurements, there may inevitably be data that is not satisfactory, such as being significantly higher or lower than expected. This could be due to measurement errors, electrical issues, or other unknown reasons. That is, the actual PSF function corresponding to one or more points, or corresponding to one or more power supplies, may be significantly higher or lower than the expected value, or significantly lower or higher than the value of the mirror point.

[0054] For example, suppose that in the crystal index, X has rows 0-46 (47 rows), and Y has rows 0-46 (47 rows). Ideally, X1Y1, X1Y63, X45Y1, and X45Y45 are corner pieces at the same mirror position at the four corners, and their measured data should be quite similar. However, in reality, the data for X1Y63, X45Y1, and X45Y45 are significantly lower or higher than the data for X1Y1. This application does not delve into the reasons for this phenomenon, because even if such a reason exists, a true PSF curve is still needed. Therefore, a solution is provided (besides re-measuring): replacing the measured data with simulated PSF curve data. The simulated PSF curve data is selected from the following: the expected value of the PSF curve data collected at that location, the estimated value based on data from 4-16 adjacent locations, or the measured value of a previously collected PSF curve data for that location. Another simple method is to directly replace the measured change curve of its mirror crystal block, or to replace it after mirroring (for example, to replace the data of X1Y1 with X1Y63, X45Y1, and X45Y45, flipping the X-axis correspondence, flipping the Y-axis correspondence, or flipping both the X and Y-axis correspondences).

[0055] Example 3

[0056] For those skilled in the art, the method of this application is somewhat abstract. To facilitate understanding, the specific implementation method is explained below. In this embodiment, a series of schematic diagrams are used to illustrate the method and help those who read this application to further understand it. However, the information in the related diagrams is explained with reference to specific conditions and should not be regarded as a necessary limitation of this application.

[0057] See Figure 1As shown, this is an example of an actual operation method for moving the point source, such as using a 3D module with a dual Z-axis gantry motion structure, placing the point source on the X-axis. This operation method imitates a common 3D printing method, and its motion algorithm can also refer to the stroke algorithm of 3D printing, making it relatively simple to program. The motion method is as follows: the X-axis module drives the point source to move left and right in the X-axis direction, satisfying the condition of corresponding to or passing through a certain Xi position; the Z-axis dual modules drive the X-axis module to move in the Z-axis direction; and the Y-axis dual modules drive the power source and the X and Z-axis modules to move in the Y-axis direction, realizing the movement of the point source across all voxel points (Xi, Yi, Zi) in the detection space. This 3D motion module is universal and easy to procure and assemble, but this is not intended to limit the invention. The invention can use any other 3D motion module that can achieve the same function, as long as it can independently drive the point source to move in the x, y, and z directions. The 3D module 2, for example, includes an X-axis module 21, a Y-axis module 22, and a Z-axis module 23.

[0058] In practice, the typical method of making a point source is to place the point source in the center of a square material. The size of the square material can be equal to or slightly larger than the crystal size. For example, the size can be selected from the following: 1.5*1.5*1.5, 1.5*1.5*3, 1.5*1.5*4, 1.7*1.7*1.7, 2*2*2, 2*2*3, 2*2*4, 2.5*2.5*2.5, 2.5*2.5*3, 2.5*2.5*4.

[0059] See Figure 2 This section provides a side view diagram of the source scanning at different heights (layers in different Z directions). The scanning FOV range (4) can be set as needed, for example, between 5-95mm, 8-12mm, 6-20mm, etc. The single-line scanning speed is between 5-150s / line, which can be adjusted based on drug activity; for example, the actual scanning speed could be 100s / line or 150s / line. The scanning interval is set to the source size; for example, if the source size is 1.7*1.7*1.7mm / cm, the scanning interval is also set to 1.7mm / cm. Data collected from each layer is stored in a separate rawdata file. Each layer is collected starting from zero; that is, before each layer is collected, Acqtime is reset to zero, and no data is collected when the termination position is reset to zero. The height (FOV) parameter information for each layer is also recorded for calculation reference. The single-layer scanning range is, for example, between 30*20mm and 160*160mm. Collimator 3 is, for example, a collimator grid or array.

[0060] Figure 3A flowchart illustrating an example scanning method is provided. After initiating a scan, the SPECT device is first initialized, specifically by setting the time window, initializing the ADC, setting AcqTime to zero, enabling DAQ output, initializing the three motors, resetting all motors, and setting the corresponding block actions according to the scan range as initialization operations. Within each layer, AcqTime is set to zero, DAQ output begins, data reception begins, data is stored in a raw file, the line feed start time is printed, the line feed motor is controlled to move and perform a line feed, the line feed end time is printed, and the line movement motor is executed to perform line-by-line scanning. After completion, it is determined whether it is the last line. If not, the process of "printing the line feed start time, controlling the line feed motor to move and perform a line feed, printing the line feed end time, and executing the line movement motor to perform line-by-line scanning" is repeated until all lines are completed. Then, it is determined whether it is the last layer. If it is, the process ends; otherwise, the control layer motors are moved, and the source and motor are moved to the next layer simultaneously. The above operations are repeated from "within each layer, ..." until the process ends.

[0061] Figure 4 The distribution of count values ​​at position X42Y22 along the z-axis for layers 0 and 20 is shown. It can be seen that the signal strength is significantly stronger at layer 20. The phantom dimensions are, for example, 1.7*1.7*1.7, 2*2*2, or 2*2*4. Single-row acquisition durations are, for example, between 10-30 seconds, such as 11.23s, 12.55s, 15.56s, 20.78s, and 18.77s. Time slice lengths are, for example, between 0.2-2 seconds, such as 1 second.

[0062] Figure 5The diagram shows a top-down view of a scan within a single layer, where, for example, the square source scan starts from the lower right corner and ends at the upper right corner. The settings here include, for example, an 82*82mm collimator, a 1.7*1.7mm grid size, and a 2*2*4mm square source size. As an example, the position of the crystal X42Y22 on the scanning plane is shown, with ΔS = (46-42)*1.7+5 = 11.8mm and ΔD = (46-22)*1.7+6 = 46.8mm. The position coordinates are, for example, (…). The positional distribution of the X42Y22 crystal count values ​​on the scanning plane (11.8, 46.8) shows that the maximum index is (5, 30) and (7, 31), corresponding to the positions as follows: (5, 30): a = 5 * 2 = 10 mm, b = 100 - (30 + 1) * 1.7 = 47.3 mm, (10 mm, 47.3 mm); (7, 31): a = 7 * 2 = 14 mm, b = 100 - (31 + 1) * 1.7 = 45.6 mm, (14 mm, 45.6 mm). The above description is merely an embodiment of the present invention; common knowledge regarding the specific structure and characteristics of the solution is not described in detail here.

[0063] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A method for obtaining a true PSF function, characterized in that: 1) For a grid-type SPECT detector equipped with a collimator, a square source having a cross-sectional area not less than that of the detector crystal grid is configured; the grid-type SPECT detector has X*Y grids; 2) Keep the SPECT detector in the detection state, make the square source in a layer within the Z layers in the Z-axis direction, set its translation speed, keep the translation speed moving in the XY plane, move the square source in a path that traverses each detector grid at least once and emits radiation, and the detector records all radiation information and its acquisition time; For other Z-1 layers, change the source to the corresponding height and repeat the above operation to obtain the radiation information and its acquisition time; 3) For a layer within the Z layers in the Z-axis direction, for the radiation information obtained in step (2) and its acquisition time, according to the movement path of the square source and the time nodes when it moves into or out of a detector grid during movement, the radiation information data is segmented, and for each grid, the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction are obtained when the square source is located on the current detector grid; The same operation is performed for the other X*Y-1 detector grids in the layer to obtain the corresponding PSF change curves in the X-axis direction and the Y-axis direction; 4) Repeat the same operation as step (3) for other Z-1 layers to obtain the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction corresponding to each detector grid in each layer; 5) Combine all PSF change curve data from steps (3) and (4) to obtain the true PSF accumulation function of the corresponding SPECT detector.

2. A SPECT imaging method, characterized in that: For current SPECT detection equipment, the real PSF function acquisition method as described in claim 1 is used to reconstruct the original radioactivity distribution in combination with the projection data obtained from a single detection, and then obtain an image that can be displayed.

3. A method for obtaining a true PSF function, characterized in that: 1) For a grid-type SPECT detector equipped with a collimator, a square source having a cross-sectional area not less than that of the detector crystal grid is configured; the grid-type SPECT detector has X*Y grids; 2) Keep the SPECT detector in the detection state, make the square source in a layer within the Z layers in the Z-axis direction, set its translation speed, keep the translation speed moving in the XY plane, move the square source in a path that traverses each detector grid at least once and emits radiation, and the detector records all radiation information and its acquisition time; For other Z-1 layers, change the source to the corresponding height and repeat the above operation to obtain the radiation information and its acquisition time; 3) For a layer within the Z layers in the Z-axis direction, for the radiation information obtained in step (2) and its acquisition time, according to the movement path of the square source and the time nodes when it moves into or out of a detector grid during movement, the radiation information data is segmented, and for each grid, the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction are obtained when the square source is located on the current detector grid; The same operation is performed for the other X*Y-1 detector grids in the layer to obtain the corresponding PSF change curves in the X-axis direction and the Y-axis direction; 4) Repeat the same operation as step (3) for other Z-1 layers to obtain the PSF change curve in the X-axis direction and the PSF change curve in the Y-axis direction corresponding to each detector grid in each layer; 5) Collect all the PSF change curve data of steps (3) and (4), and replace one or more unreasonable PSF change curve data with simulated PSF change curve data; 6) Merge all the PSF change curve data collected and replaced in step (5) to obtain the real PSF accumulation function of the corresponding SPECT detector.

4. A method for obtaining a true PSF function as claimed in claim 3, characterized in that: The simulated PSF change curve data is selected from the following data: a mathematical expectation value of the PSF change curve data collected for the position, a calculated value based on the data of 4-16 adjacent groups of positions, or a measured value of the PSF change curve data collected for the position last time; The rules for judging unreasonable PSF change curve data are as follows: for a certain PSF change curve, its mathematical expectation value is obtained according to the measured data of the PSF change curves at the adjacent 4-16 positions. If the PSF change curve data collected at this position is compared with the mathematical expectation value and is generally lower by A% or higher by more than A%, it is considered unreasonable, A∈[1,10].

5. A SPECT imaging method, characterized in that: For current SPECT detection equipment, the real PSF function acquisition method as described in claim 3 or 4 is used to reconstruct the original radioactivity distribution in combination with the projection data obtained from a single detection, and then obtain an image that can be displayed.