WO2023056634A1 - Pet参数成像方法、装置、电子设备及可读存储介质 - Google Patents

Pet参数成像方法、装置、电子设备及可读存储介质 Download PDF

Info

Publication number
WO2023056634A1
WO2023056634A1 PCT/CN2021/122865 CN2021122865W WO2023056634A1 WO 2023056634 A1 WO2023056634 A1 WO 2023056634A1 CN 2021122865 W CN2021122865 W CN 2021122865W WO 2023056634 A1 WO2023056634 A1 WO 2023056634A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
pet
voxel
target
feature
Prior art date
Application number
PCT/CN2021/122865
Other languages
English (en)
French (fr)
Inventor
胡战利
郑海荣
梁栋
杨永峰
刘新
毛鑫
Original Assignee
深圳先进技术研究院
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳先进技术研究院 filed Critical 深圳先进技术研究院
Priority to PCT/CN2021/122865 priority Critical patent/WO2023056634A1/zh
Publication of WO2023056634A1 publication Critical patent/WO2023056634A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T11/00—2D [Two Dimensional] image generation

Definitions

  • the present application relates to the technical field of Positron Emission Tomography (PET) imaging, and in particular to a PET parametric imaging method, device, electronic equipment and readable storage medium.
  • PET Positron Emission Tomography
  • PET parametric imaging is based on dynamic PET reconstruction, followed by kinetic model fitting to obtain the changing parameters of the radiotracer.
  • Physiological changes in organisms can be directly quantified by these parameters. For example, blood flow, glucose metabolism, and receptor binding, among others.
  • voxel intensity information in Magnetic Resonance Imaging (MRI) with high spatial resolution is usually used as prior information to regularize PET reconstruction.
  • MRI Magnetic Resonance Imaging
  • the emergence of the system has greatly improved the image quality obtained by the existing MRI-guided PET parametric imaging method.
  • more attention is paid to the voxel intensity information in the spatial domain of MRI, and the aspects considered are relatively simple, so there is still a lot of room for improvement in the image quality obtained by the parametric imaging method.
  • the present application discloses a PET parametric imaging method, device, electronic equipment and readable storage medium, in order to solve the problem that the current parametric imaging method considers relatively single factors, resulting in a large room for improvement in the imaging quality of the imaging image.
  • the embodiment of the present application discloses a PET parametric imaging method, including: acquiring the PET dynamic data and the first image of the target object; acquiring the prior image of the first image, wherein the prior image includes texture A characteristic image, and/or a voxel intensity image; acquiring a kernel matrix of the first image according to the prior image; determining a target parameter image according to the kernel matrix and the PET dynamic data.
  • the embodiment of the present application discloses a PET parameter imaging device, including: a first acquisition module, configured to acquire PET dynamic data and a first image of a target object; a second acquisition module, configured to acquire the first image A priori image, wherein the priori image includes a texture feature image, and/or a voxel intensity image; a third acquisition module, configured to acquire the kernel matrix of the first image according to the priori image; the determination module, It is used for determining a target parameter image according to the kernel matrix and the PET dynamic data.
  • the embodiment of the present application discloses an electronic device, including a processor, a memory, and a program or instruction stored in the memory and operable on the processor, and the program or instruction is executed by the processor When executed, the steps of the method described in the first aspect are realized.
  • the embodiment of the present application discloses a readable storage medium, where a program or an instruction is stored on the readable storage medium, and when the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented.
  • the embodiment of the present application discloses a PET parameter imaging method, by acquiring the PET dynamic data and the first image of the target object; acquiring the prior image of the first image, wherein the prior image includes a texture feature image, and/or Or a voxel intensity image; according to the prior image, obtain the kernel matrix of the first image; according to the kernel matrix and the PET dynamic data, determine the target parameter image.
  • the texture feature image and voxel intensity image can be processed to obtain a kernel matrix containing voxel intensity information and texture feature information, and the kernel matrix and PET
  • the dynamic data is processed to obtain a new PET parameter image, that is, the target parameter image, so that the imaging quality of the target parameter image is greatly improved under the condition that the tracer dose remains unchanged.
  • the embodiment of the present application discloses a PET parametric imaging method, which can solve the problem that the imaging quality of the imaging image still has a large room for improvement due to relatively single factors considered by the current parametric imaging method.
  • FIG. 1 is a schematic flow diagram of a PET parameter imaging method disclosed in an embodiment of the present application
  • FIG. 2 is a schematic structural diagram of a PET parameter imaging device disclosed in an embodiment of the present application.
  • FIG. 3 is a schematic structural diagram of an electronic device disclosed in an embodiment of the present application.
  • FIG. 1 is a schematic flowchart of a PET parametric imaging method, which can be executed by an electronic device, in other words, the method can be executed by software or hardware installed in the electronic device.
  • the electronic device may be a PET/MRI system or the like.
  • a PET parameter imaging method disclosed in the embodiment of the present application may include the following steps:
  • S110 Acquire PET dynamic data and a first image of the target object.
  • the target object can be scanned, so that the PET dynamic data and the first image of the target object can be obtained.
  • the first image can be an MRI image.
  • the first image may also be a medical image such as a computer tomography (Computed Tomography, CT) image, which is not specifically limited in this embodiment of the present application.
  • S120 Acquire a priori image of the first image.
  • the prior image includes a texture feature image, and/or a voxel intensity image.
  • correlation processing is performed on the first image to obtain a priori image including a texture feature image and/or a voxel intensity image.
  • S130 Acquire a kernel matrix of the first image according to the prior image.
  • the texture feature image and/or the voxel intensity image in the prior image it can be processed to obtain a new kernel matrix, that is, the kernel matrix of the first image.
  • the prior image can be processed so that a kernel matrix can be obtained from the texture feature image and the voxel intensity image; when the prior image includes the texture feature image
  • the texture feature image can be processed, so that a kernel matrix can be obtained from the texture feature image; when the prior image includes a voxel intensity image, the voxel intensity image can be processed, so that the volume Get a kernel matrix from the voxel intensity image.
  • S140 Determine a target parameter image according to the kernel matrix and the PET dynamic data.
  • the kernel matrix and the PET dynamic data obtained by scanning are processed, and the PET dynamic data can be reconstructed according to the processed kernel matrix and the general PET dynamic data, so as to obtain a new PET parameter image, that is, a target parameter image.
  • a new PET parameter image that is, a target parameter image.
  • the prior image includes texture feature image and voxel intensity image
  • a kernel matrix containing texture feature image information and voxel intensity image information can be obtained, the kernel matrix is processed, and combined with the processed PET dynamic data , so that the target parameter image with better imaging quality can be obtained.
  • the embodiment of the present application discloses a PET parameter imaging method, by acquiring the PET dynamic data and the first image of the target object; acquiring the prior image of the first image, wherein the prior image includes a texture feature image, and/or Or a voxel intensity image; according to the prior image, obtain the kernel matrix of the first image; according to the kernel matrix and the PET dynamic data, determine the target parameter image.
  • the texture feature image and voxel intensity image can be processed to obtain a kernel matrix containing voxel intensity information and texture feature information, and the kernel matrix and PET
  • the dynamic data is processed to obtain a new PET parameter image, that is, the target parameter image, so that the imaging quality of the target parameter image is greatly improved under the condition that the tracer dose remains unchanged.
  • the embodiment of the present application discloses a PET parametric imaging method, which can solve the problem that the imaging quality of the imaging image still has a large room for improvement due to relatively single factors considered by the current parametric imaging method.
  • the acquiring the prior image of the first image may include:
  • Step 1 Filter the target module according to the target module centered on the target voxel in the first image to obtain a second image.
  • the first image includes multiple target voxels.
  • the target module is filtered, so that the filtered image, that is, the second image can be obtained, and the first image includes a plurality of target voxels, so that Multiple target modules respectively corresponding to multiple target voxels need to be filtered, so that multiple second images respectively corresponding to multiple target modules can be obtained.
  • Step 2 Transform the second image into a frequency domain by Fourier transform to obtain a third image.
  • the second image is convoluted through Fourier transform, so that the second image can be transformed into the frequency domain to obtain the third image.
  • Step 3 Obtain the autocorrelation feature value of the target module according to the third image and the coordinates of the third image.
  • the autocorrelation feature value is a feature value corresponding to the target voxel in the texture feature image, and a plurality of the autocorrelation feature values constitute the texture feature image.
  • calculate the autocorrelation feature value of the target module and use it as the feature value corresponding to the target voxel in the texture feature image, so that all target objects in the first image can be obtained
  • the autocorrelation feature value corresponding to the pixel, so that the texture feature image can be obtained according to multiple autocorrelation feature values.
  • Step 4 Acquire the prior image according to the texture feature image and the pre-acquired voxel intensity image.
  • the texture feature image and voxel intensity image can be used as the prior image, through which the kernel matrix can be obtained from the prior image, so that according to the kernel Matrix and PET dynamic data to obtain target parameter images.
  • the first image can be an MRI image, centering on a single voxel i in the MRI image (i.e. the target voxel described above), and acquiring a module with a window size of 5 ⁇ 5 (i.e. The above-mentioned target module), the module can be filtered by the Garbor function, wherein, the implementation formula of filtering the module by the two-dimensional Garbor function g(x,y) is as follows:
  • g(x, y) represents the function image of the second image obtained after filtering the target module
  • ⁇ represents the space length ratio
  • ⁇ represents the standard deviation of the Gaussian factor
  • ⁇ indicates the direction of the filter
  • the value of ⁇ / ⁇ determines the spatial frequency bandwidth.
  • ⁇ / ⁇ can take the default value of 0.56
  • the Gabor function is used to convolve the MRI image, so that the second image can be transformed into the frequency domain to obtain the third image.
  • the implementation formula for convolving the Gabor function with the MRI image is as follows:
  • the symbol * represents convolution
  • I'(u, v) represents the image after convolution (that is, the third image mentioned above)
  • I represents the MRI image
  • the function F(x) represents Fourier transform.
  • the obtained third image calculate the autocorrelation eigenvalue of the selected module, and use it as the eigenvalue of the corresponding voxel i in the texture feature image, so that the autocorrelation eigenvalues of all target voxels can be obtained, so that the texture feature can be obtained image.
  • the implementation formula for calculating the autocorrelation eigenvalue of the texture feature image is as follows:
  • the method for obtaining texture feature images in the present application is not limited to the Gabor method, and methods based on other transformations can also be used: such as wavelet transform; based on statistical methods: Gray-Level Run-Length Matrix (Gray-Level Run-Length Matrix, GLRLM), local binary pattern (Local Binary Pattern, LBP), etc.; even including the texture information extraction method based on deep learning, which is not specifically limited in the embodiment of the present application.
  • the acquiring the kernel matrix of the first image according to the prior image may include:
  • the texture feature image and the voxel intensity image can be used as the prior image at the same time, and the kernel matrix of the first image is obtained from the prior image through the K-Nearest Neighbor (KNN) algorithm.
  • KNN K-Nearest Neighbor
  • the window size can be selected as 7 ⁇ 7, and the nearest neighbor value is 20, so that a new kernel function can be formed by using the first voxel feature vector of the texture feature image and the second voxel feature vector of the voxel intensity image , the implementation formula of the new kernel function is as follows:
  • the texture feature image and the voxel intensity image can be used as the prior image at the same time, so that the kernel matrix can be obtained according to the prior image.
  • the determination of the target parameter image according to the kernel matrix and the PET dynamic data may include:
  • Step 1 Obtain PET sinusoidal data according to the PET dynamic data, wherein the PET sinusoidal data includes noise information.
  • PET sinusoidal data including noise information can be obtained.
  • the PET sinusoidal data may include 24 time frames: 4x20s, 4x40s, 4x60s, 4x180s, and 8x300s.
  • Step 2 Determine the target parameter image according to the PET sinusoidal data and the kernel matrix.
  • the target parametric image can be determined by a direct parametric kernel reconstruction method.
  • the implementation formula of the direct parametric kernel reconstruction method is as follows:
  • ⁇ represents the parameter image to be obtained.
  • ⁇ ⁇ represents the coefficient of ⁇
  • n represents the number of iterations
  • the superscript T represents the transposition of the matrix
  • A represents the parameter coefficient of the dynamic model
  • P represents the system matrix
  • 1 M represents the vector of all 1s
  • y represents the PET sine data
  • ⁇ represents background events, including random events and scattering events
  • this application can also use the Maximum A Posteriori (Maximum A Posteriori, MAP) imaging method.
  • the proposed method is not limited to the linear irreversible Patlak model. Other models (such as Logan Graphic model) and other reversible models are also applicable.
  • the method proposed in this application is not limited to PET parametric imaging and PET static imaging and dynamic imaging, and can also be applied to imaging of other modality data (including MR, CT, etc.).
  • the embodiment of the present application discloses a PET parameter imaging method.
  • a new kernel matrix can be obtained from the prior image, so that the new kernel matrix contains
  • the voxel intensity image and the texture feature image in the prior image are taken into account, so that the factors considered in the parameter imaging are more comprehensive, so that the imaging quality of the target parameter image can be further improved when the tracer remains unchanged.
  • the embodiment of the present application discloses a PET parameter imaging device. As shown in FIG. Module 230 and determination module 240, wherein,
  • the first acquisition module 210 is configured to acquire the PET dynamic data and the first image of the target object.
  • the second obtaining module 220 is configured to obtain a prior image of the first image, wherein the prior image includes a texture feature image and/or a voxel intensity image.
  • the third acquiring module 230 is configured to acquire the kernel matrix of the first image according to the prior image.
  • a determining module 240 configured to determine a target parameter image according to the kernel matrix and the PET dynamic data.
  • the second acquiring module 220 is used to:
  • the target module is filtered to obtain a second image, wherein the first image includes a plurality of target voxels; through Fourier transform, Transforming the second image into the frequency domain to obtain a third image; obtaining an autocorrelation feature value of the target module according to the third image and the coordinates of the third image, wherein the autocorrelation feature value is The feature value corresponding to the target voxel in the texture feature image, a plurality of the autocorrelation feature values constitute the texture feature image; according to the texture feature image and the pre-acquired voxel intensity image, obtain The prior image.
  • the third acquiring module 230 is used to:
  • the determination module 240 is used to:
  • the PET parameter imaging apparatus 200 in the embodiment of the present application may be a device, or a component, an integrated circuit, or a chip in a terminal.
  • the device may be a mobile electronic device or a non-mobile electronic device.
  • the mobile electronic device may be a PET/MRI system, etc.
  • the non-mobile electronic device may be a server, a network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in this embodiment of the present application.
  • Network Attached Storage Network Attached Storage
  • the PET parameter imaging device 200 in the embodiment of the present application may be a device with an operating system.
  • the operating system may be an Android (Android) operating system, an ios operating system, or other possible operating systems, which are not specifically limited in this embodiment of the present application.
  • the PET parameter imaging apparatus 200 provided in the embodiment of the present application can implement various processes implemented in the method embodiment in FIG. 1 , and details are not repeated here to avoid repetition.
  • the embodiment of the present application also provides an electronic device 300, including a processor 301, a memory 302, and a program or instruction stored in the memory 302 and operable on the processor 301.
  • the electronic devices in the embodiments of the present application include the above-mentioned mobile electronic devices and non-mobile electronic devices.
  • the embodiment of the present application also provides a readable storage medium, on which a program or instruction is stored, and when the program or instruction is executed by a processor, each process of the above-mentioned PET parameter imaging method embodiment can be achieved, and the same Technical effects, in order to avoid repetition, will not be repeated here.
  • a readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a magnetic disk or an optical disk, and the like.
  • ROM computer read-only memory
  • RAM random access memory
  • magnetic disk or an optical disk and the like.
  • the term “comprising”, “comprising” or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or device. Without further limitations, an element defined by the phrase “comprising a " does not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising that element.
  • the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved. Functions are performed, for example, the described methods may be performed in an order different from that described, and various steps may also be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

一种PET参数成像方法、装置、电子设备及可读存储介质，所述方法包括：获取目标对象的PET动态数据和第一图像（S110）；获取所述第一图像的先验图像（S120），其中，所述先验图像包括纹理特征图像，和/或体素强度图像；根据先验图像，获取所述第一图像的核矩阵（S130）；根据所述核矩阵和所述PET动态数据，确定目标参数图像（S140）。

Description

PET参数成像方法、装置、电子设备及可读存储介质 技术领域
本申请涉及正电子发射断层显像(Positron Emission Tomography，PET)成像技术领域，尤其涉及一种PET参数成像方法、装置、电子设备及可读存储介质。
背景技术
PET参数成像是基于动态PET重建，然后进行动力学模型拟合，得到放射性示踪剂的变化参数。通过这些参数可以直接的量化生物体内的生理学变化。例如，血流量、葡萄糖代谢和受体结合等。
为了提高PET参数成像的质量，通常会利用具有高空间分辨的核磁共振成像(Magnetic Resonance Imaging，MRI)中的体素强度信息作为先验信息对PET重建进行正则化，随着PET/MRI一体化系统的出现，使得已有的MRI引导PET参数成像的方法得到的图像质量获得了巨大的提高。然而，目前的成像方法中，更多的还是关注MRI的空间域中体素强度信息，考虑的方面较为单一，从而通过参数成像的方法得到的图像质量还有很大的进步空间。
发明内容
本申请公开一种PET参数成像方法、装置、电子设备及可读存储介质，以解决目前的参数成像方法考虑的因素较为单一导致成像图像的成像质量还有较大进步空间的问题。
为了解决上述问题，本申请采用下述技术方案：
第一方面，本申请实施例公开一种PET参数成像方法，包括：获取目标对象的PET动态数据和第一图像；获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；根据先验图像，获取所述第一图像的核矩阵；根据所述核矩阵和所述PET动态数据，确定目标参数图像。
第二方面，本申请实施例公开一种PET参数成像装置，包括：第一获取模块，用于获取目标对象的PET动态数据和第一图像；第二获取模块，用于获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；第三获取模块，用于根据先验图像，获取所述第一图像的核矩阵；确定模块，用于根据所述核矩阵和所述PET动态数据，确定目标参数图像。
第三方面，本申请实施例公开一种电子设备，包括处理器，存储器及存储在所述存储器上并可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第四方面，本申请实施例公开一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如第一方面所述的方法的步骤。
本申请实施例公开本申请采用的技术方案能够达到以下有益效果：
本申请实施例公开一种PET参数成像方法，通过获取目标对象的PET动态数据和第一图像；获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；根据先验图像，获取所述第一图像的核矩阵；根据所述核矩阵和所述PET动态数据，确定目标参数图像。使得在先验图像包括纹理特征图像和体素强度图像的情况下，可以对纹理特征图像和体素强度图像进行处理，得到包含体素强度信息和纹理特征信息的核矩阵，对核矩阵和PET动态数据进行处理，从而可以得到新的PET参数图像，即目标参数图像，从而在示踪剂剂量不变的情况下，目标参数图像的成像质量得到较大地提高。也就是说，本申请实施例公开一种PET参数成像方法，可以解决目前的参数成像方法考虑的因素较为单一导致成像图像的成像质量还有较大进步空间的问题。
附图说明
图1为本申请实施例公开的一种PET参数成像方法的流程示意图；
图2为本申请实施例公开的一种PET参数成像装置的结构示意图；
图3为本申请实施例公开的一种电子设备的结构示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
本申请的说明书和权利要求书中的术语“第一”、“第二”等是用于区别类似的对象，而不用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施，且“第一”、“第二”等所区分的对象通常为一类，并不限定对象的个数，例如第一对象可以是一个，也可以是多个。此外，说明书以及权利要求中“和/或”表示所连接对象的至少其中之一，字符“/”，一般表示前后关联对象是一种“或”的关系。
下面结合附图，通过具体的实施例及其应用场景对本申请实施例提供一种PET参数成像方法、装置、电子设备及可读存储介质进行详细地说明。
图1为一种PET参数成像方法的流程示意图，该方法可以由电子设备执行，换言之，该方法可以由安装在电子设备的软件或硬件来执行。示例性的，电子设备可以为PET/MRI系统等。如图1所示，本申请实施例公开的一种PET参数成像方法可以包括以下步骤：
S110：获取目标对象的PET动态数据和第一图像。
通过PET/MRI系统，可以对目标对象进行扫描，从而可以得到目标对象的PET动态数据和第一图像，此时，第一图像可以是MRI图像。当然，第一图像也可以是电子计算机断层扫描(Computed Tomography，CT)图像等医学图像，本申请实施例对此不作具体限制。
S120：获取所述第一图像的先验图像。
其中，所述先验图像包括纹理特征图像，和/或体素强度图像。在得到第一图像之后，对第一图像进行相关处理，可以得到包括纹理特征图像，和/或体素强度图像的先验图像。
S130：根据先验图像，获取所述第一图像的核矩阵。
根据先验图像中的纹理特征图像，和/或体素强度图像，可以对其进行处理，得到新的核矩阵，即第一图像的核矩阵。在先验图像包括纹理特征图像和体素强度图像的情况下，可以对先验图像进行处理，从而可以从纹理特征图像和体素强度图像中获取一个核矩阵；在先验图像包括纹理特征图像的情况下，可以对纹理特征图像进行处理，从而可以从纹理特征图像中获取一个核矩阵；在先验图像包括体素强度图像的情况下，可以对体素强度图像进行处理，从而可以从体素强度图像中获取一个核矩阵。
S140：根据所述核矩阵和所述PET动态数据，确定目标参数图像。
对核矩阵和扫描得到的PET动态数据进行处理，根据处理后的核矩阵和通PET动态数据，可以对PET动态数据进行重建，从而得到新的PET参数图像，即目标参数图像。在先验图像包括纹理特征图像和体素强度图像的情况下，可以获得一个包含纹理特征图像信息和体素强度图像信息的核矩阵，对该核矩阵进行处理，并结合处理后的PET动态数据，从而可以得到成像质量较好的目标参数图像。
本申请实施例公开一种PET参数成像方法，通过获取目标对象的PET动态数据和第一图像；获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；根据先验图像，获取所述第一图像的核矩阵；根据所述核矩阵和所述PET动态数据，确定目标参数图像。使得在先验图像包括纹理特征图像和体素强度图像的情况下，可以对纹理特征图像和体素强度图像进行处理，得到包含体素强度信息和纹理特征信息的核矩阵，对核矩阵和PET动态数据进行处理，从而可以得到新的PET参数图像，即目标参数图 像，从而在示踪剂剂量不变的情况下，目标参数图像的成像质量得到较大地提高。也就是说，本申请实施例公开一种PET参数成像方法，可以解决目前的参数成像方法考虑的因素较为单一导致成像图像的成像质量还有较大进步空间的问题。
一种可以实现的方式中，所述获取所述第一图像的先验图像，可以包括：
步骤1：根据所述第一图像中以目标体素为中心的目标模块，对所述目标模块进行滤波得到第二图像。
其中，所述第一图像中包括多个目标体素。通过获取到的以第一图像中目标体素为中心的目标模块，对目标模块进行滤波处理，从而可以得到滤波后的图像即第二图像，而第一图像中包括多个目标体素，从而要对多个目标体素分别对应的多个目标模块进行滤波处理，从而可以得到多个目标模块分别对应的多个第二图像。
步骤2：通过傅里叶变换，将所述第二图像变换到频域得到第三图像。
根据第二图像和第一图像，通过傅里叶变换对第二图像进行卷积，从而可以将第二图像变换到频域得到第三图像。
步骤3：根据所述第三图像和所述第三图像的坐标，获取所述目标模块的自相关特征值。
其中，所述自相关特征值为所述纹理特征图像中与所述目标体素对应的特征值，多个所述自相关特征值构成所述纹理特征图像。基于得到的第三图像和第三图像的坐标，计算目标模块的自相关特征值，并将其作为纹理特征图像中与目标体素对应的特征值，从而可以求得第一图像中所有目标体素对应的自相关特征值，从而根据多个自相关特征值可以得到纹理特征图像。
步骤4：根据所述纹理特征图像和预先获取的所述体素强度图像，获取所述先验图像。
根据获取的纹理特征图像和预先获取的体素强度图像，可以将纹理特征图像和体素强度图像作为先验图像，通过该先验图像，可以从先验图像中获取核 矩阵，从而可以根据核矩阵和PET动态数据，得到目标参数图像。
一种具体的实施例中，第一图像可以为MRI图像，以MRI图像中的单个体素i(即上文所述的目标体素)为中心，获取窗口大小为5×5的模块(即上文所述的目标模块)，通过Garbor函数可以对该模块进行滤波，其中，通过二维Garbor函数g(x,y)对该模块进行滤波的实现公式如下：
Figure PCTCN2021122865-appb-000001
Figure PCTCN2021122865-appb-000002
Figure PCTCN2021122865-appb-000003
其中，g(x,y)表示对目标模块滤波后得到第二图像的函数图像，γ表示空间长度比，σ表示高斯因子标准差，
Figure PCTCN2021122865-appb-000004
表示该模块在滤波前的坐标，θ表示滤波器方向、
Figure PCTCN2021122865-appb-000005
表示相位偏移，σ/λ的值决定了空间频率带宽。可选地，σ/λ可以取默认值0.56，角度参数
Figure PCTCN2021122865-appb-000006
可选地，可以取
Figure PCTCN2021122865-appb-000007
从而可以使得函数偶对称，角度参数θ∈[0，π)，可选地，θ＝π/4。之后，通过Gabor函数与MRI图像进行卷积，从而可以将第二图像变换到频域得到第三图像，通过Gabor函数与MRI图像进行卷积的实现公式如下：
I'(u,v)＝F(||(I*g)(x,y)|| 2)，
其中，符号*表示卷积，I'(u,v)表示卷积后的图像(即上文所述的第三图像)，I表示MRI图像，函数F(x)表示傅里叶变换。
根据得到的第三图像，计算所取模块的自相关特征值，将其作为纹理特征图像中对应体素i的特征值，从而可以得到所有目标体素的自相关特征值，从而可以得到纹理特征图像。计算纹理特征图像的自相关特征值的实现公式如下：
Figure PCTCN2021122865-appb-000008
需要说明的是，本申请中获取纹理特征图像的方法不局限于Gabor方法，还可使用基于其他变换的方法：如小波变换；基于统计学方法：灰度游程长度矩阵(Gray-Level Run-Length Matrix，GLRLM)、局部二值模式(Local Binary  Pattern，LBP)等；甚至包括基于深度学习的纹理信息提取方式，本申请实施例对此不作具体限制。
一种可以实现的方式中，所述根据先验图像，获取所述第一图像的核矩阵，可以包括：
获取所述纹理特征图像中的第一体素特征向量和所述体素强度图像中的第二体素特征向量；根据所述第一体素特征向量和所述第二体素特征向量，获取所述核矩阵。
具体的，可以将纹理特征图像和体素强度图像同时作为先验图像，通过K最邻近(K-Nearest Neighbor，KNN)算法从先验图像中求取第一图像的核矩阵。根据获取到的纹理特征图像中的第一体素特征向量和体素强度图像中的第二体素特征向量，计算核矩阵，在以体素i为中心的窗口内，选取制定数量与此体素强度最为接近的体素j，并求取两者的欧几里得距离。本申请中，可以选取窗口大小为7×7，最邻近数值为20，从而通过纹理特征图像的第一体素特征向量和体素强度图像的第二体素特征向量，可以构成新的核函数，构成新的核函数的实现公式如下：
Figure PCTCN2021122865-appb-000009
Figure PCTCN2021122865-appb-000010
其中，
Figure PCTCN2021122865-appb-000011
和
Figure PCTCN2021122865-appb-000012
分别表示体素i和j的特征向量，Auto表示纹理特征图像，σ表示图像的标准差。
通过这种方式，可以将纹理特征图像和体素强度图像同时作为先验图像，从而根据该先验图像，可以得到核矩阵。
一种可以实现的方式中，所述根据所述核矩阵和所述PET动态数据，确定目标参数图像，可以包括：
步骤1：根据所述PET动态数据，获取PET正弦数据，其中，所述PET正弦数据包括噪声信息。
对PET动态数据进行衰减校正、添加泊松噪声和20％的背景时间，从而可以得到包括噪声信息的PET正弦数据。该PET正弦数据可以包括24个时间帧：4×20s，4×40s，4×60s，4×180s，和8×300s。
步骤2：根据所述PET正弦数据和所述核矩阵，确定所述目标参数图像。
利用PET正弦数据和第一图像的核矩阵，可以通过直接参数核重建方法确定目标参数图像。直接参数核重建方法的实现公式如下：
Figure PCTCN2021122865-appb-000013
θ＝Kα θ；
其中，θ表示待求的参数图像。α θ表示θ的系数，n表示迭代次数，上标T表示矩阵的转置，A表示动力学模型的参数系数，P标识系统矩阵，1 M标识全为1的向量，y表示PET正弦数据，γ表示背景事件，包括随机事件和散射事件，
Figure PCTCN2021122865-appb-000014
表示Kronecker积。通过直接参数核重建方法，可以降低目标参数图像的噪声，从而可以提高目标参数图像的成像质量。
除了核重建方法，本申请也可以采用最大后验概率(Maximum A Posteriori，MAP)成像方法，此外，在参数成像中，所提出的方法不局限于线性不可逆的Patlak模型，其他的模型(如Logan图模型)以及其他可逆的模型也适用，本申请所提出的方法不局限于PET参数成像和PET的静态成像、动态成像，也可以适用于其它模态数据的成像(包括MR、CT等)。
本申请实施例公开一种PET参数成像方法，通过将纹理特征图像和体素强度图像同时作为先验图像，从而可以从该先验图像中求取新的核矩阵，从而新的核矩阵中包含了先验图像中的体素强度图像和纹理特征图像，从而在参数成像时考虑的因素较为全面，从而在示踪剂保持不变的情况下，目标参数图像的成像质量可以得到进一步地提高。
基于上文所述的PET参数成像方法，本申请实施例公开一种PET参数成像装置，如图2所示，该装置200可以包括：第一获取模块210、第二获取模 块220、第三获取模块230和确定模块240，其中，
第一获取模块210，用于获取目标对象的PET动态数据和第一图像。
第二获取模块220，用于获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像。
第三获取模块230，用于根据先验图像，获取所述第一图像的核矩阵。
确定模块240，用于根据所述核矩阵和所述PET动态数据，确定目标参数图像。
一种可以实现的方式中，所述第二获取模块220用于：
根据所述第一图像中以目标体素为中心的目标模块，对所述目标模块进行滤波得到第二图像，其中，所述第一图像中包括多个目标体素；通过傅里叶变换，将所述第二图像变换到频域得到第三图像；根据所述第三图像和所述第三图像的坐标，获取所述目标模块的自相关特征值，其中，所述自相关特征值为所述纹理特征图像中与所述目标体素对应的特征值，多个所述自相关特征值构成所述纹理特征图像；根据所述纹理特征图像和预先获取的所述体素强度图像，获取所述先验图像。
一种可以实现的方式中，所述第三获取模块230用于：
获取所述纹理特征图像中的第一体素特征向量和所述体素强度图像中的第二体素特征向量；根据所述第一体素特征向量和所述第二体素特征向量，获取所述核矩阵。
一种可以实现的方式中，所述确定模块240用于：
根据所述PET动态数据，获取PET正弦数据，其中，所述PET正弦数据包括噪声信息；根据所述PET正弦数据和所述核矩阵，确定所述目标参数图像。
本申请实施例中的PET参数成像装置200可以是装置，也可以是终端中的部件、集成电路、或芯片。该装置可以是移动电子设备，也可以为非移动电子设备。示例性的，移动电子设备可以为PET/MRI系统等，非移动电子设备可以为服务器、网络附属存储器(Network Attached Storage，NAS)等，本申请 实施例不作具体限定。
本申请实施例中的PET参数成像装置200可以为具有操作系统的装置。该操作系统可以为安卓(Android)操作系统，可以为ios操作系统，还可以为其他可能的操作系统，本申请实施例不作具体限定。
本申请实施例提供的PET参数成像装置200能够实现图1的方法实施例中实现的各个过程，为避免重复，这里不再赘述。
可选的，如图3所示，本申请实施例还提供一种电子设备300，包括处理器301，存储器302，存储在存储器302并可在处理器301上运行的程序或指令，该程序或指令被处理器301执行时实现上述PET参数成像方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
需要说明的是，本申请实施例中的电子设备包括上述所述的移动电子设备和非移动电子设备。
本申请实施例还提供一种可读存储介质，可读存储介质上存储有程序或指令，该程序或指令被处理器执行时实现上述PET参数成像方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
其中，处理器为上述实施例中所述的电子设备中的处理器。可读存储介质，包括计算机可读存储介质，如计算机只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。此外，需要指出的是，本申请实施方式中的方法和装置的范围不限按示出或讨论的顺序来执行功能，还可包括根据所涉及的功能按基本同时的方式或按相反的顺序来执行功能，例如，可以按不同于所 描述的次序来执行所描述的方法，并且还可以添加、省去、或组合各种步骤。另外，参照某些示例所描述的特征可在其他示例中被组合。
本申请上文实施例中重点描述的是各个实施例之间的不同，各个实施例之间不同的优化特征只要不矛盾，均可以组合形成更优的实施例，考虑到行文简洁，在此则不再赘述。
以上所述仅为本申请的实施例而已，并不用于限制本申请。对于本领域技术人员来说，本申请可以有各种更改和变化。凡在本申请的精神和原理之内所作的任何修改、等同替换、改进等，均应包含在本申请的权利要求范围之内。

Claims (10)

  1. 一种PET参数成像方法，包括：
    获取目标对象的PET动态数据和第一图像；
    获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；
    根据先验图像，获取所述第一图像的核矩阵；
    根据所述核矩阵和所述PET动态数据，确定目标参数图像。
  2. 根据权利要求1所述的方法，其中，所述获取所述第一图像的先验图像，包括：
    根据所述第一图像中以目标体素为中心的目标模块，对所述目标模块进行滤波得到第二图像，其中，所述第一图像中包括多个目标体素；
    通过傅里叶变换，将所述第二图像变换到频域得到第三图像；
    根据所述第三图像和所述第三图像的坐标，获取所述目标模块的自相关特征值，其中，所述自相关特征值为所述纹理特征图像中与所述目标体素对应的特征值，多个所述自相关特征值构成所述纹理特征图像；
    根据所述纹理特征图像和预先获取的所述体素强度图像，获取所述先验图像。
  3. 根据权利要求1所述的方法，其中，所述根据先验图像，获取所述第一图像的核矩阵，包括：
    获取所述纹理特征图像中的第一体素特征向量和所述体素强度图像中的第二体素特征向量；
    根据所述第一体素特征向量和所述第二体素特征向量，获取所述核矩阵。
  4. 根据权利要求1所述的方法，其中，所述根据所述核矩阵和所述PET 动态数据，确定目标参数图像，包括：
    根据所述PET动态数据，获取PET正弦数据，其中，所述PET正弦数据包括噪声信息；
    根据所述PET正弦数据和所述核矩阵，确定所述目标参数图像。
  5. 一种PET参数成像装置，包括：
    第一获取模块，用于获取目标对象的PET动态数据和第一图像；
    第二获取模块，用于获取所述第一图像的先验图像，其中，所述先验图像包括纹理特征图像，和/或体素强度图像；
    第三获取模块，用于根据先验图像，获取所述第一图像的核矩阵；
    确定模块，用于根据所述核矩阵和所述PET动态数据，确定目标参数图像。
  6. 根据权利要求5所述的装置，其中，所述第二获取模块用于：
    根据所述第一图像中以目标体素为中心的目标模块，对所述目标模块进行滤波得到第二图像，其中，所述第一图像中包括多个目标体素；通过傅里叶变换，将所述第二图像变换到频域得到第三图像；根据所述第三图像和所述第三图像的坐标，获取所述目标模块的自相关特征值，其中，所述自相关特征值为所述纹理特征图像中与所述目标体素对应的特征值，多个所述自相关特征值构成所述纹理特征图像；根据所述纹理特征图像和预先获取的所述体素强度图像，获取所述先验图像。
  7. 根据权利要求5所述的装置，其中，所述第三获取模块用于：
    获取所述纹理特征图像中的第一体素特征向量和所述体素强度图像中的第二体素特征向量；根据所述第一体素特征向量和所述第二体素特征向量，获取所述核矩阵。
  8. 根据权利要求5所述的装置，其中，所述确定模块用于：
    根据所述PET动态数据，获取PET正弦数据，其中，所述PET正弦数据包括噪声信息；根据所述PET正弦数据和所述核矩阵，确定所述目标参数图像。
  9. 一种电子设备，包括处理器，存储器及存储在所述存储器上并可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求1-4任一项所述的方法的步骤。
  10. 一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如权利要求1-4任一项所述的方法的步骤。
PCT/CN2021/122865 2021-10-09 2021-10-09 Pet参数成像方法、装置、电子设备及可读存储介质 WO2023056634A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/122865 WO2023056634A1 (zh) 2021-10-09 2021-10-09 Pet参数成像方法、装置、电子设备及可读存储介质

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2021/122865 WO2023056634A1 (zh) 2021-10-09 2021-10-09 Pet参数成像方法、装置、电子设备及可读存储介质

Publications (1)

Publication Number Publication Date
WO2023056634A1 true WO2023056634A1 (zh) 2023-04-13

Family

ID=85803854

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/122865 WO2023056634A1 (zh) 2021-10-09 2021-10-09 Pet参数成像方法、装置、电子设备及可读存储介质

Country Status (1)

Country Link
WO (1) WO2023056634A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016067254A1 (en) * 2014-10-30 2016-05-06 Koninklijke Philips N.V. Texture analysis map for image data
CN109308728A (zh) * 2018-10-25 2019-02-05 上海联影医疗科技有限公司 正电子发射型计算机断层扫描图像处理方法及装置
US20200320753A1 (en) * 2019-04-03 2020-10-08 Uih America, Inc. Systems and methods for positron emission tomography image reconstruction
CN112488952A (zh) * 2020-12-07 2021-03-12 深圳先进技术研究院 Pet图像的重建方法及重建终端、计算机可读存储介质
CN112614200A (zh) * 2020-12-18 2021-04-06 深圳先进技术研究院 Pet图像的重建方法、装置及设备

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016067254A1 (en) * 2014-10-30 2016-05-06 Koninklijke Philips N.V. Texture analysis map for image data
CN109308728A (zh) * 2018-10-25 2019-02-05 上海联影医疗科技有限公司 正电子发射型计算机断层扫描图像处理方法及装置
US20200320753A1 (en) * 2019-04-03 2020-10-08 Uih America, Inc. Systems and methods for positron emission tomography image reconstruction
CN112488952A (zh) * 2020-12-07 2021-03-12 深圳先进技术研究院 Pet图像的重建方法及重建终端、计算机可读存储介质
CN112614200A (zh) * 2020-12-18 2021-04-06 深圳先进技术研究院 Pet图像的重建方法、装置及设备

Similar Documents

Publication Publication Date Title
CN107527359B (zh) 一种pet图像重建方法及pet成像设备
US11633146B2 (en) Automated co-registration of prostate MRI data
EP2380132B1 (en) Denoising medical images
Uzunova et al. Memory-efficient GAN-based domain translation of high resolution 3D medical images
Chen et al. CT metal artifact reduction method based on improved image segmentation and sinogram in‐painting
US20130202177A1 (en) Non-linear resolution reduction for medical imagery
JP6275826B2 (ja) ノイズ除去再構成画像データエッジ改善
EP3501006B1 (en) Feature-based image processing using feature images extracted from different iterations
Takenaga et al. Modulation transfer function measurement of CT images by use of a circular edge method with a logistic curve-fitting technique
Zhu et al. Residual dense network for medical magnetic resonance images super-resolution
US20220058798A1 (en) System and method for providing stroke lesion segmentation using conditional generative adversarial networks
Budhiraja et al. An efficient image denoising scheme for higher noise levels using spatial domain filters
JP5815573B2 (ja) 機能的画像データ強調方法及びエンハンサ
Huang et al. Context-aware, reference-free local motion metric for CBCT deformable motion compensation
CN108038840A (zh) 一种图像处理方法、装置、图像处理设备及存储介质
Nguyen et al. Medical image denoising via optimal implementation of non-local means on hybrid parallel architecture
WO2023056634A1 (zh) Pet参数成像方法、装置、电子设备及可读存储介质
WO2011100575A2 (en) Systems, methods and computer readable storage mediums storing instructions for applying multiscale bilateral filtering to magnetic resonance (mr) images
Dutta et al. PET point spread function modeling and image deblurring using a PET/MRI joint entropy prior
EP3658031B1 (en) Motion compensated cardiac valve reconstruction
Xu et al. Tensor scale: An analytic approach with efficient computation and applications
CN113920214B (zh) Pet参数成像方法、装置、电子设备及可读存储介质
Park et al. Performance evaluation of improved median-modified Wiener filter with segmentation method to improve resolution in computed tomographic images
Alrefaya et al. A novel adaptive probabilistic nonlinear denoising approach for enhancing PET data sinogram
da Silveira Souza et al. Detector of 3-D salient points based on the dual-tree complex wavelet transform for the positioning of hippocampi meshes in magnetic resonance images

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21959724

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21959724

Country of ref document: EP

Kind code of ref document: A1