A method, system, electronic device and computer readable storage medium for ROI region positioning for PET device image quality assessment

Abstract

The application provides an ROI region positioning method for PET device image quality evaluation, comprising the following steps: obtaining an initial three-dimensional image and a standard reference image; using a rigid registration algorithm to obtain a transformation matrix of the initial three-dimensional image to the standard reference image, so that the initial three-dimensional image is transformed according to the transformation matrix to obtain a registered three-dimensional image; obtaining a registered intermediate tomographic image in the registered three-dimensional image, obtaining a corresponding ball ROI region of any one of the plurality of filling balls in the registered intermediate tomographic image, and obtaining a plurality of corresponding background ROI regions of any one of the plurality of filling balls in the registered three-dimensional image. After the above technical scheme is used, the central tomographic image passing through the centers of all the filling balls can be obtained through the image registration operation, so that the required ROI region for quality evaluation can be accurately positioned, and the dependence on the positioning accuracy of the IEC phantom is reduced.

Description

Technical Field

[0001] This invention relates to the field of nuclear medicine imaging equipment, and more particularly to a method, system, electronic device, and computer-readable storage medium for ROI region localization in PET device image quality assessment. Background Technology

[0002] To measure whether a PET device's performance meets IEC (International Electrotechnical Commission) industry standards, an IEC phantom with several infill spheres is typically used to scan the PET device under evaluation, obtaining a PET image of the IEC phantom. The contrast percentage and background change rate of each infill sphere within the PET image are used as evaluation indicators of the PET device's image quality. Before obtaining these evaluation indicators, it is necessary to obtain the ROI (Region of Interest) areas of the infill spheres to calculate the event count for each ROI area.

[0003] Existing methods for obtaining the ROI region of spheres typically involve positioning the IEC phantom with high precision before scanning to ensure that the central tomographic image of the PET image passes through the center of each infill sphere, followed by manual annotation on this central tomographic image. This ROI region positioning method is highly dependent on the pre-scan positioning. If the IEC phantom is misaligned within the PET device, it cannot be guaranteed that the central tomographic image passes through the center of all infill spheres, leading to inaccurate positioning of the sphere's ROI region and affecting the quality assessment results. Alternatively, locating the tomographic image of each infill sphere with the highest activity individually would make obtaining the background ROI region corresponding to each infill sphere excessively complex, time-consuming, and labor-intensive. Summary of the Invention

[0004] In order to overcome the above-mentioned technical defects, the purpose of this invention is to provide a method, system, device and medium for ROI region localization for image quality assessment of PET devices, which is used to accurately locate the position of the filling sphere through image registration.

[0005] This invention discloses a method for ROI region localization for image quality assessment of PET devices, characterized by comprising the following steps: Acquire an initial 3D image and a standard reference image, wherein the initial 3D image is a PET image obtained by scanning an IEC phantom with multiple infill spheres inside in the PET device, the centers of the multiple infill spheres are located on the same plane and their diameters are all different, and the standard reference image is a simulated PET image of the IEC phantom; A rigid registration algorithm is used to obtain the transformation matrix from the initial 3D image to the standard reference image, and the initial 3D image is transformed according to the transformation matrix to obtain a registered 3D image; In the registered 3D image, an intermediate tomographic image is obtained, and in the intermediate tomographic image, a sphere ROI region corresponding to any one of the plurality of filled spheres is obtained, wherein any one of the plurality of filled spheres has the same diameter as its corresponding sphere ROI region; In the registered 3D image, obtain multiple background ROI regions corresponding to any one of the multiple filled spheres, wherein any one of the multiple filled spheres has the same diameter as the multiple background ROI regions corresponding to it.

[0006] Preferably, before registering the initial 3D image, the method further includes the following steps: The initial 3D image and the standard reference image are normalized to make the pixel size of the initial 3D image the same as that of the standard reference image, and the 3D data dimension of the initial 3D image the same as that of the standard reference image.

[0007] Preferably, before registering the initial 3D image, the following steps are further included: A sample tomographic image is obtained from the initial three-dimensional image. The sample tomographic image is a tomographic image of the multiple filled spheres. Based on the peak position of the X-axis projection of the sample tomographic image, it is determined whether the initial three-dimensional image is in a horizontally reversed state. When it is determined that the initial 3D image is in a horizontally reversed state, the initial 3D image is horizontally mirrored and flipped. If it is determined that the initial 3D image is not in a horizontally reversed state, the initial 3D image is maintained.

[0008] Preferably, obtaining the spherical ROI region of the plurality of filled spheres includes: The phantom center of the IEC phantom is obtained based on the registered intermediate tomographic image; Based on the relative distance between the center of any of the plurality of infilled spheres and the center of the IEC phantom, and the diameter of any of the plurality of infilled spheres, obtain the ROI region of any of the plurality of infilled spheres in the registered intermediate tomographic image.

[0009] Preferably, obtaining the spherical ROI region of the plurality of filled spheres further includes: The edges of the multiple spherical ROI regions are traversed point by point. Based on the ratio of the coverage area of ​​any point in its corresponding pixel to the area of ​​the complete pixel, the value weight of the pixel corresponding to that point is obtained, forming a mask for the multiple spherical ROI regions, which is used to optimize the edges of the multiple spherical ROI regions.

[0010] Preferably, obtaining multiple background ROI regions includes: In the registered 3D image, multiple extended tomographic images are acquired forward and backward, centered on the registered intermediate tomographic image. In the registered intermediate tomographic image and the plurality of extended tomographic images, a plurality of background ROI regions corresponding to any one of the plurality of filled spheres are obtained, wherein any one of the plurality of filled spheres and its corresponding plurality of background ROI regions have the same diameter, and the plurality of background ROI regions and the plurality of spherical ROI regions do not overlap spatially.

[0011] This invention also discloses a ROI region localization system for image quality assessment of PET devices, characterized in that it includes: The image acquisition module is used to acquire an initial three-dimensional image and a standard reference image. The initial three-dimensional image is a PET image obtained by scanning an IEC phantom with multiple infill spheres inside it in the PET device. The centers of the multiple infill spheres are located on the same plane and their diameters are all different. The standard reference image is a simulated PET image of the IEC phantom. The image registration module is used to obtain the transformation matrix from the initial 3D image to the standard reference image using a rigid registration algorithm, so that the initial 3D image is transformed according to the transformation matrix to obtain a registered 3D image; The ROI positioning module is used to obtain a registration intermediate tomographic image in the registration 3D image, and to obtain the ROI region of any one of the plurality of filled spheres in the registration intermediate tomographic image, wherein any one of the plurality of filled spheres has the same diameter as its corresponding ROI region. The ROI positioning module obtains multiple background ROI regions corresponding to any one of the multiple filled spheres in the registered 3D image, wherein any one of the multiple filled spheres has the same diameter as its corresponding multiple background ROI regions.

[0012] The present invention also discloses an electronic device, which includes a memory storing computer-executable instructions and a processor, wherein when the instructions are executed by the processor, the electronic device performs the aforementioned ROI region positioning method.

[0013] The present invention also discloses a computer-readable storage medium storing a computer program thereon, characterized in that, when the computer program is run on a computer, it causes the computer to execute the aforementioned ROI region positioning method.

[0014] Compared with existing technologies, the above technical solution has the following advantages: it reduces the dependence on the positioning accuracy of the IEC phantom. Regardless of whether the IEC phantom is tilted or offset, the central tomographic image passing through the center of all filled spheres can be obtained through image registration to accurately locate the ROI area required for quality assessment. Attached Figure Description

[0015] Figure 1 A flowchart illustrating a method for ROI region localization in PET device image quality assessment provided by the present invention. Figure 2 A schematic diagram of the sample tomographic image and X-axis projection provided by the present invention; Figure 3 This is a schematic diagram of the registration intermediate tomographic image provided by the present invention; Figure 4 This is a schematic diagram of the spherical ROI region mask provided by the present invention; Figure 5 This is a schematic diagram of the spherical ROI region mask provided by the present invention; Figure 6 This is a schematic diagram of the recovery contrast data and the change curve of the background change rate provided by the present invention. Detailed Implementation

[0016] The advantages of the present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments.

[0017] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0018] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0019] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0020] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0021] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0022] In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the convenience of the description of the invention and have no specific meaning in themselves. Therefore, "module" and "part" can be used interchangeably.

[0023] like Figure 1 As shown, this invention discloses a method for ROI region localization for image quality assessment of PET devices, characterized by comprising the following steps: S100. Obtain an initial three-dimensional image and a standard reference image, wherein the initial three-dimensional image is a PET image obtained by scanning an IEC phantom with multiple infilled spheres inside in the PET device, the centers of the multiple infilled spheres are located on the same plane and their diameters are all different, and the standard reference image is a simulated PET image of the IEC phantom; S200. A rigid registration algorithm is used to obtain the transformation matrix from the initial three-dimensional image to the standard reference image, and the initial three-dimensional image is transformed according to the transformation matrix to obtain a registered three-dimensional image. S300. Obtain a registration intermediate tomographic image from the registration three-dimensional image, and obtain the ROI region corresponding to any one of the plurality of filled spheres from the registration intermediate tomographic image, wherein any one of the plurality of filled spheres has the same diameter as its corresponding ROI region. S400. Obtain multiple background ROI regions corresponding to any one of the multiple filled spheres in the registered 3D image, wherein any one of the multiple filled spheres has the same diameter as the multiple background ROI regions corresponding to it.

[0024] Specifically, the IEC phantom includes a background cavity and multiple infill spheres placed within the cavity. The centers of these spheres are located on the same plane, and their diameters are all unequal. Typically, four infill spheres with diameters of 10mm, 13mm, 17mm, and 22mm are selected as hot spheres, and two infill spheres with diameters of 28mm and 37mm are selected as cold spheres. A first radioactive tracer is injected into the four hot spheres, and non-radioactive purified water is injected into the two cold spheres. A second radioactive tracer is injected into the background cavity. The radioactivity of the first radioactive tracer is 4 to 8 times the radioactivity concentration of the second radioactive tracer. The IEC phantom is placed on the scanning bed of the PET device, and a complete single-bed scan is performed to reconstruct an initial 3D image. A rigid registration algorithm is used to register the initial 3D image with a reference 3D image, obtaining the ROI region of each infill sphere and the background ROI region under ideal conditions. Finally, the quality analysis indicators of the PET device are calculated based on the event counts of the sphere ROI regions and the background ROI regions.

[0025] The essence of the rigid registration algorithm is to find a transformation matrix that, when applied to the image to be registered, yields a transformed image that minimizes the difference between the transformed image and the reference image. This difference can be defined by a distance function, such as the square of the Euclidean distance, the mean absolute difference, or the sum of squared errors. In this invention, the rigid registration algorithm is used to obtain the transformation matrix from the initial 3D image to the standard reference image. This transformation matrix is ​​then applied to the initial 3D image to obtain the registered 3D image. Ultimately, the geometric center of the registered 3D image is located at the exact center of the scanning field of view of the PET device, and the central slice of the registered 3D image is the slice containing the centers of all the filled spheres. (Reference) Figure 3 As shown, Figure 3 (a) is the intermediate tomographic image of the initial three-dimensional image. At this time, due to the positioning error of the IEC phantom, the intermediate tomographic image of the PET image failed to pass through the center of all the filled spheres. Figure 3 (b) is an intermediate tomographic image of the standard reference image. Figure 3(c) is the registration intermediate tomographic image of the registered three-dimensional image. After the initial three-dimensional image is registered into the registered three-dimensional image, the registration center tomographic image of the registered three-dimensional image passes through the center of all the filled spheres.

[0026] Commonly used rigid registration algorithms include the ICP (Iterative Closest Point) algorithm in point set registration techniques or the 3D rigid registration algorithm in ITK. Taking the ICP algorithm as an example, assuming the initial 3D image is: The standard reference image is: The registration process involves each point in the initial 3D image P. Find its nearest point in the standard reference image X Using Euclidean distance as the metric, the objective function for registration is: in, It is a quaternion variable. To convert the quaternion into a rotation matrix form, T is the translation matrix. Solving the ICP is essentially solving for the minimum of the objective function described above. Each iteration calculates the parameters that minimize the objective function. After convergence, T and T are returned. This is the transformation matrix for the final registration.

[0027] The main process for solving the objective function is as follows: Assume the centroids of the initial 3D image P and the standard reference image X are respectively and The initial 3D image and the standard reference image are centered, and then the covariance matrix is ​​calculated: Calculate the covariance matrix and perform SVD decomposition. in, It is the covariance matrix The singular value matrix, and These are orthogonal matrices.

[0028] Calculate the rotation matrix and translation matrix Calculate the objective function, and stop iterating if it is less than the threshold.

[0029] Preferably, before registering the initial 3D image, the following steps are also included: The initial 3D image and the standard reference image are normalized to make the pixel size of the initial 3D image the same as that of the standard reference image, and the 3D data dimension of the initial 3D image the same as that of the standard reference image.

[0030] Specifically, firstly, the pixel dimensions of the initial 3D image and the standard reference image are read. Using the image with the smaller pixel dimension as the benchmark, interpolation is performed on the image with the larger pixel dimension. Interpolation methods such as nearest neighbor interpolation, multiple spline interpolation, 3D cubic interpolation, and 3D linear interpolation can be used to unify the pixel dimensions of the initial 3D image and the standard reference image. Then, the data dimensions of the initial 3D image and the standard reference image are compared. The image with the larger data dimension is used as the volume center as the benchmark point for segmentation, thus unifying the data dimensions of the initial 3D image and the standard reference image.

[0031] Preferably, before registering the initial 3D image, the following steps are also included: A sample tomographic image is obtained from the initial three-dimensional image. The sample tomographic image is a tomographic image of the multiple filled spheres. Based on the peak position of the X-axis projection of the sample tomographic image, it is determined whether the initial three-dimensional image is in a horizontally reversed state. When it is determined that the initial 3D image is in a horizontally reversed state, the initial 3D image is horizontally mirrored and flipped. If it is determined that the initial 3D image is not in a horizontally reversed state, the initial 3D image is maintained.

[0032] Specifically, such as Figure 2 As shown, the projection of a tomographic image of multiple filled spheres (not necessarily passing through the center of all the spheres) onto the X-axis is obtained, and the peak position is observed. For example... Figure 2 (a) The projection peak is on the left, indicating that the cold ball is on the right. The placement of the IEC phantom has not been horizontally reversed, and no horizontal adjustment is needed to the initial 3D image. Figure 2 (b) The projection peak is on the right, indicating that the cold ball is on the left. The IEC phantom is placed horizontally and reversed. The initial three-dimensional image is horizontally and reversed. The initial three-dimensional image needs to be horizontally mirrored and flipped before the registration step is performed.

[0033] Preferably, obtaining the spherical ROI region of the plurality of filled spheres includes: The phantom center of the IEC phantom is obtained based on the registered intermediate tomographic image; Based on the relative distance between the center of any of the plurality of infilled spheres and the center of the IEC phantom, and the diameter of any of the plurality of infilled spheres, obtain the ROI region of any of the plurality of infilled spheres in the registered intermediate tomographic image.

[0034] Specifically, the positions of the six spheres in the IEC phantom are determined, with 37mm and 28mm being cold spheres and the rest being hot spheres. The six spheres are distributed around the center of the cross section of the registered intermediate tomographic image, with an angle of 60 degrees between any two spheres, and the center of each infilled sphere is 57.2mm away from the center of the phantom. Then, based on the distance between the center of any infilled sphere and the center of the phantom, as well as the diameter of the sphere, the ROI region of the sphere on the registered intermediate tomographic image can be obtained.

[0035] Preferably, obtaining the spherical ROI region of the plurality of filled spheres further includes: The edges of the multiple spherical ROI regions are traversed point by point. Based on the ratio of the coverage area of ​​any point in its corresponding pixel to the area of ​​the complete pixel, the value weight of the pixel corresponding to that point is obtained, forming a mask for the multiple spherical ROI regions, which is used to optimize the edges of the multiple spherical ROI regions.

[0036] Specifically, the spherical ROI region is a circle with the same diameter as its corresponding filling sphere on the registered intermediate tomographic image. Since a pixel is a square with equal sides, traditional methods typically directly assign values ​​to pixels covered by the circle when locating the ROI region. This leads to excessive edge artifacts and affects the quality assessment results. Therefore, this invention optimizes ROI region location as follows: when any point on the edge of the ROI region falls on a pixel (but cannot cover the entire pixel), the ratio of the area covered by that point in its corresponding pixel to the area of ​​the complete pixel is used as the pixel's value weight. This results in pixels closer to the edge of the ROI region having lower values, reducing edge artifacts. The improved mask for a spherical ROI region is as follows: Figure 4 As shown, the masks for the last 6 spherical ROI regions are as follows: Figure 5 As shown.

[0037] Preferably, multiple background ROI regions are obtained, including: In the registered 3D image, multiple extended tomographic images are acquired forward and backward, centered on the registered intermediate tomographic image. In the registered intermediate tomographic image and the plurality of extended tomographic images, a plurality of background ROI regions corresponding to any one of the plurality of filled spheres are obtained, wherein any one of the plurality of filled spheres and its corresponding plurality of background ROI regions have the same diameter, and the plurality of background ROI regions and the plurality of spherical ROI regions do not overlap spatially.

[0038] Specifically, on the registered intermediate tomographic image, 12 background ROI regions corresponding to any infilled sphere are obtained. Each background ROI region is 15mm from the edge of the IEC phantom, and each background ROI region is at least 15mm away from the ROI region of the corresponding infilled sphere. For example, for a 37mm cold sphere, 12 37mm background ROI regions are obtained on the registered intermediate tomographic image. Each background ROI region is 15mm from the edge of the IEC phantom, and each background ROI region is at least 15mm away from the ROI region of the corresponding infilled sphere. The above steps are repeated for infilled spheres of 10mm, 13mm, 17mm, 22mm, and 28mm, obtaining 12 background ROI regions for each infilled sphere.

[0039] In the registered 3D images, two extended tomographic images are acquired forward and two backward from the registered intermediate tomographic image as the center, resulting in four extended tomographic images. For each infill sphere, 12 background ROI regions are obtained in each extended tomographic image. Therefore, each infill sphere obtains a total of 60 background ROI regions of the same diameter across the five tomographic images. Through these steps, the ROI regions of spheres with the same diameter and the count data within multiple background ROI regions are obtained, and finally, the image quality evaluation index of the PET device is calculated.

[0040] The image quality assessment metrics for PET devices include recovery contrast and background change rate.

[0041] Recovery contrast of each hot ball j In the formula: The average count within the ROI region of the hot sphere j; This represents the average count within the background ROI region of the hotspot j; The radioactivity concentration of the first radioactive tracer; This represents the radioactivity concentration of the second radioactive tracer.

[0042] The restored contrast of each cold ball j In the formula: The average count within the ROI region of the cold sphere j; This represents the average count of the background ROI region of the cold ball j.

[0043] The rate of change of the background of sphere j In the formula: Let be the standard deviation of the background count of the ROI for sphere j.

[0044] Ultimately, we can obtain the recovery contrast data and background change rate data as a function of the infill sphere diameter, and plot the change curves based on these data, such as... Figure 6 As shown, it can be used to evaluate the image quality of a PET device.

[0045] This invention also discloses a ROI region localization system for image quality assessment of PET devices, characterized in that it includes: The image acquisition module is used to acquire an initial three-dimensional image and a standard reference image. The initial three-dimensional image is a PET image obtained by scanning an IEC phantom with multiple infill spheres inside it in the PET device. The centers of the multiple infill spheres are located on the same plane and their diameters are all different. The standard reference image is a simulated PET image of the IEC phantom. The image registration module uses a rigid registration algorithm to obtain the transformation matrix from the initial 3D image to the standard reference image, and transforms the initial 3D image according to the transformation matrix to obtain a registered 3D image. The ROI positioning module is used to obtain a registration intermediate tomographic image in the registration 3D image, and to obtain the ROI region of any one of the plurality of filled spheres in the registration intermediate tomographic image, wherein any one of the plurality of filled spheres has the same diameter as its corresponding ROI region. The ROI positioning module obtains multiple background ROI regions corresponding to any one of the multiple filled spheres in the registered 3D image, wherein any one of the multiple filled spheres has the same diameter as its corresponding multiple background ROI regions.

[0046] The present invention also discloses an electronic device, which includes a memory storing computer-executable instructions and a processor, wherein when the instructions are executed by the processor, the electronic device performs the aforementioned ROI region positioning method.

[0047] The present invention also discloses a computer-readable storage medium storing a computer program thereon, characterized in that, when the computer program is run on a computer, it causes the computer to execute the aforementioned ROI region positioning method.

[0048] It should be noted that the embodiments of the present invention have better implementability and are not intended to limit the present invention in any way. Any person skilled in the art may use the above-disclosed technical content to change or modify it into equivalent effective embodiments. However, any modifications or equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of the technical solution of the present invention.

Claims

1. A method for ROI region localization for image quality assessment of PET devices, characterized in that, Includes the following steps: Acquire an initial 3D image and a standard reference image, wherein the initial 3D image is a PET image obtained by scanning an IEC phantom with multiple infill spheres inside in the PET device, the centers of the multiple infill spheres are located on the same plane and their diameters are all different, and the standard reference image is a simulated PET image of the IEC phantom; A rigid registration algorithm is used to obtain the transformation matrix from the initial 3D image to the standard reference image, and the initial 3D image is transformed according to the transformation matrix to obtain a registered 3D image; In the registered 3D image, an intermediate tomographic image is obtained, and in the intermediate tomographic image, a sphere ROI region corresponding to any one of the plurality of filled spheres is obtained, wherein any one of the plurality of filled spheres has the same diameter as its corresponding sphere ROI region; In the registered 3D image, obtain multiple background ROI regions corresponding to any one of the multiple filled spheres, wherein any one of the multiple filled spheres and its corresponding multiple background ROI regions have the same diameter; Before registering the initial 3D image, the following steps are also included: A sample tomographic image is obtained from the initial three-dimensional image. The sample tomographic image is a tomographic image of the multiple filled spheres. Based on the peak position of the X-axis projection of the sample tomographic image, it is determined whether the initial three-dimensional image is in a horizontally reversed state. When it is determined that the initial 3D image is in a horizontally reversed state, the initial 3D image is horizontally mirrored and flipped. If it is determined that the initial 3D image is not in a horizontally reversed state, the initial 3D image is maintained.

2. The ROI region positioning method according to claim 1, characterized in that, Before registering the initial 3D image, the following steps are also included: The initial 3D image and the standard reference image are normalized to make the pixel size of the initial 3D image the same as that of the standard reference image, and the 3D data dimension of the initial 3D image the same as that of the standard reference image.

3. The ROI region positioning method according to claim 1, characterized in that, Obtaining the spherical ROI regions of the plurality of filled spheres, including: The phantom center of the IEC phantom is obtained based on the registered intermediate tomographic image; Based on the relative distance between the center of any of the plurality of infilled spheres and the center of the IEC phantom, and the diameter of any of the plurality of infilled spheres, obtain the ROI region of any of the plurality of infilled spheres in the registered intermediate tomographic image.

4. The ROI region positioning method according to claim 3, characterized in that, Obtaining the spherical ROI region of the plurality of filled spheres further includes: The edges of the multiple spherical ROI regions are traversed point by point. Based on the ratio of the coverage area of ​​any point in its corresponding pixel to the area of ​​the complete pixel, the value weight of the pixel corresponding to that point is obtained, forming a mask for the multiple spherical ROI regions, which is used to optimize the edges of the multiple spherical ROI regions.

5. The ROI region positioning method according to claim 1, characterized in that, Multiple baseline ROI regions were obtained, including: In the registered 3D image, multiple extended tomographic images are acquired forward and backward, centered on the registered intermediate tomographic image. In the registered intermediate tomographic image and the plurality of extended tomographic images, a plurality of background ROI regions corresponding to any one of the plurality of filled spheres are obtained, wherein any one of the plurality of filled spheres and its corresponding plurality of background ROI regions have the same diameter, and the plurality of background ROI regions and the plurality of spherical ROI regions do not overlap spatially.

6. A ROI region localization system for image quality assessment of a PET device, characterized in that, include: The image acquisition module is used to acquire an initial three-dimensional image and a standard reference image. The initial three-dimensional image is a PET image obtained by scanning an IEC phantom with multiple infill spheres inside it in the PET device. The centers of the multiple infill spheres are located on the same plane and their diameters are all different. The standard reference image is a simulated PET image of the IEC phantom. The image registration module is used to obtain the transformation matrix from the initial 3D image to the standard reference image using a rigid registration algorithm, so that the initial 3D image is transformed according to the transformation matrix to obtain a registered 3D image; The ROI positioning module is used to obtain a registration intermediate tomographic image in the registration 3D image, and to obtain the ROI region of any one of the plurality of filled spheres in the registration intermediate tomographic image, wherein any one of the plurality of filled spheres has the same diameter as its corresponding ROI region. The ROI positioning module obtains multiple background ROI regions corresponding to any one of the multiple filled spheres in the registered 3D image, wherein any one of the multiple filled spheres has the same diameter as the multiple background ROI regions corresponding to it. Before registering the initial 3D image, the following steps are also included: A sample tomographic image is obtained from the initial three-dimensional image. The sample tomographic image is a tomographic image of the multiple filled spheres. Based on the peak position of the X-axis projection of the sample tomographic image, it is determined whether the initial three-dimensional image is in a horizontally reversed state. When it is determined that the initial 3D image is in a horizontally reversed state, the initial 3D image is horizontally mirrored and flipped. If it is determined that the initial 3D image is not in a horizontally reversed state, the initial 3D image is maintained.

7. An electronic device comprising a memory storing computer-executable instructions and a processor, wherein when the instructions are executed by the processor, the electronic device performs the ROI region localization method according to any one of claims 1-5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is run on a computer, it causes the computer to perform the ROI region location method according to any one of claims 1-5.

Images