WO2022165882A1 - 一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质 - Google Patents

一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2022165882A1
WO2022165882A1 PCT/CN2021/077972 CN2021077972W WO2022165882A1 WO 2022165882 A1 WO2022165882 A1 WO 2022165882A1 CN 2021077972 W CN2021077972 W CN 2021077972W WO 2022165882 A1 WO2022165882 A1 WO 2022165882A1
Authority
WO
WIPO (PCT)
Prior art keywords
thin layer
cardiomyocyte
voxel position
inclination angle
expression form
Prior art date
Application number
PCT/CN2021/077972
Other languages
English (en)
French (fr)
Inventor
李洪莹
Original Assignee
四川大学
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 四川大学 filed Critical 四川大学
Publication of WO2022165882A1 publication Critical patent/WO2022165882A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00Three dimensional [3D] modelling, e.g. data description of 3D objects
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H30/00ICT specially adapted for the handling or processing of medical images
    • G16H30/20ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10072Tomographic images
    • G06T2207/10088Magnetic resonance imaging [MRI]

Definitions

  • the invention belongs to the technical field of medical image processing, and in particular relates to a method, device, computer equipment and computer-readable storage medium for reconstructing a thin-layer arrangement structure of myocardial cells.
  • Cardiac muscle is a kind of muscle tissue composed of cardiomyocytes, among which, cardiomyocytes (cardiac muscle cells or cardiomyocytes) are the smallest units that constitute the atrium and ventricle. organization arrangement. Cardiomyocytes are connected by intercalated discs, which are locally directional in the myocardium and spiral in the heart wall; at the same time, cardiomyocytes have another organizational form in the myocardium: they are about four cardiomyocytes thick. Units are stacked in sheetlets.
  • Cardiomyocytes are arranged in a spiral shape, so that the heart twists during contraction and relaxation (observed from the short-axis of the heart); the cardiomyocytes are stacked into a thin layer structure so that the adjacent thin layers of the heart slide relative to each other during contraction, tending to be perpendicular to the heart wall. , the heart wall thickens to minimize the ventricular volume to maximize pumping to the arteries, while during diastole, the adjacent lamellae slide relative to each other and gradually tend to be erect, perpendicular to the short-axis plane of the heart, and the heart wall becomes thinner , to maximize ventricular blood volume.
  • the two basic organizational forms of the aforementioned cardiomyocytes are the microstructural basis for the heart to achieve its function.
  • Diffusion Tensor Imaging can obtain the diffusion characteristics of water molecules in the myocardium in a non-invasive way, and the diffusion characteristics of water molecules in the myocardium can indirectly reflect the arrangement of myocardial cells. From the Diffusion Weighted Image (DWI) output by DTI imaging, the diffusion tensor can be calculated. DTI imaging uses the form of diffusion tensor to characterize the diffusion of water in muscle tissue.
  • DTI Diffusion Tensor Imaging
  • the diffusion tensor can be decomposed into three eigenvalues ⁇ 1 , ⁇ 2 and ⁇ 3 (arranged in descending order according to the size of the eigenvalues), and the corresponding three eigenvectors v (1) , v (2) and v (3 ) (that is, the first eigenvector primary eigenvector, the second eigenvector secondary eigenvector, and the third eigenvector tertiary eigenvector are all unit vectors, and the three eigenvectors are orthogonal), where the first eigenvector v (1) is often considered to be Consistent with the orientation of the cardiomyocytes, it can be used to calculate the helix angle (HA), which quantifies the helical morphology exhibited by the arrangement of cardiomyocytes (thought to reveal myocardial necrosis).
  • HA helix angle
  • the dynamic variation range of the thin-layer structure composed of cardiomyocytes can more directly reflect the pumping capacity of the heart, that is, during the beating process of the heart, the thin layer is inclined relative to the sliding, and the larger the angle of inclination, the greater the change in the heart volume. , the stronger the blood pumping ability.
  • the existing thin layer modeling method is to use the plane composed of the eigenvectors v (1) and v (2) to represent the myocardial cell thin layer;
  • the existing thin layer inclination angle calculation method is to use only a single one-dimensional eigenvector v (2 ) to quantify the change in the angle of the thin layer is often referred to as E2A, E2 angle.
  • the purpose of the present invention is to provide a myocardial cell thin layer Reconstruction method, device, computer equipment and computer-readable storage medium of arrangement structure, by using the directed plane of double vector to represent the thin layer formed by the arrangement of cardiomyocytes through geometric algebra G3 space, and redefining in this modeling way
  • the thin-layer inclination quantification calculation method can avoid the thin-layer modeling error and simplify the thin-layer inclination quantitative calculation process, which is beneficial to quickly obtain the reconstruction results of the thin-layer arrangement of myocardial cells, and is convenient for clinical medicine based on diffusion tensor imaging.
  • Real-time reconstruction can be applied to magnetic resonance imaging instruments and diffusion magnetic resonance image analysis/processing software, which is convenient for practical application and promotion.
  • the present invention provides a method for reconstructing a myocardial cell thin layer arrangement structure, comprising:
  • the geometric algebraic expression form of the diffusion tensor everywhere in the diffusion weighted image, wherein the geometric algebraic expression form includes three eigenvalues ⁇ 1 , ⁇ 2 and ⁇ 3 of the diffusion tensor and the corresponding three unit features vectors v (1) , v (2) and v (3) , the three unit eigenvectors v (1) , v (2) and v (3) are perpendicular to each other;
  • the double vector expression form of the cardiomyocyte thin layer at voxel position d is calculated by solving the following optimization problem:
  • S represents the double vector expression form of the cardiomyocyte thin layer
  • L(S) represents the objective function with respect to the variable S
  • ⁇ 0 represents the first preset parameter
  • k is a natural number used to control the size of the neighborhood around the voxel position d
  • i is a natural number
  • ⁇ i is the same as the voxel position d +i ⁇ corresponds to the second preset parameter
  • represents a unit voxel
  • represents the norm calculation symbol
  • represents the outer multiplication
  • the corresponding thin layer inclination angle is calculated according to the following formula:
  • represents the inclination angle of the thin layer
  • * represents the scalar product operation
  • S r represents the dual vector expression form of the preset reference plane, Indicates the negated or reversed form of S r ;
  • the output shows the dual vector representations of all cardiomyocyte lamellae and the corresponding lamella tilt angles to obtain the reconstruction result of the cardiomyocyte lamella arrangement structure of the myocardium.
  • a novel reconstruction scheme that can effectively reduce the thin layer modeling error and simplify the thin layer inclination quantization calculation process is provided, that is, the cardiomyocyte arrangement is represented by the directed plane of the double vector based on the geometric algebra G 3 space.
  • the thin layer formed, and the thin layer inclination quantification calculation method is redefined under this modeling method, which can avoid the thin layer modeling error, and simplify the thin layer inclination angle quantification calculation process, which is beneficial to quickly obtain the myocardial cell thin layer arrangement structure.
  • the reconstruction result is convenient for real-time reconstruction based on diffusion tensor imaging in clinical medicine, and can be applied to magnetic resonance imaging instruments and diffusion magnetic resonance image analysis/processing software, which is convenient for practical application and promotion.
  • the method further includes:
  • a value corresponding to a preset size is assigned to the thin-layer direction similarity index at the voxel position d in the reconstruction result evaluation map, wherein the coordinate variable of the reconstruction result evaluation map is voxel Position d, the thin layer direction similarity index is used to characterize the degree of direction similarity between the cardiomyocyte thin layer at the corresponding voxel position and the cardiomyocyte thin layer at the adjacent voxel position;
  • the output shows an evaluation map of the reconstruction results.
  • the calculation formula of the inclination angle of the thin layer can be replaced by:
  • the output displays a dual vector representation of all cardiomyocyte lamellae and the corresponding lamella tilt angles, including:
  • a corresponding preset color is assigned according to the magnitude of the inclination angle of the thin layer of the cardiomyocyte thin layer.
  • the output displays a dual vector representation of all cardiomyocyte lamellae and the corresponding lamella tilt angles, including:
  • a value corresponding to a preset size is assigned to the inclination angle index at the voxel position d in the inclination angle map of the thin layer, wherein the coordinate variable of the inclination angle map of the thin layer is the volume voxel position d, the inclination size index is used to characterize the size of the included angle between the thin layer of cardiomyocytes and the preset reference plane at the corresponding voxel position;
  • the output shows the lamella tilt angle map.
  • the preset reference plane adopts the cardiac short-axis plane.
  • the present invention provides a reconstruction device for the arrangement of thin layers of cardiomyocytes, including an expression form acquisition module, a dual vector calculation module, an inclination angle calculation module and an output display module that are sequentially communicated and connected;
  • the expression form obtaining module is used to obtain the geometric algebraic expression form of the diffusion tensor everywhere in the diffusion weighted image, wherein the geometric algebraic expression form includes three eigenvalues ⁇ 1 , ⁇ 2 and ⁇ of the diffusion tensor 3 and the corresponding three unit eigenvectors v (1) , v (2) and v (3) , the three unit eigenvectors v (1) , v (2) and v (3) are perpendicular to each other;
  • the dual vector calculation module is used to calculate the dual vector representation of the cardiomyocyte thin layer at the voxel position d by solving the following optimization problem according to the geometric algebraic representation of the dispersion tensor:
  • S represents the double vector expression form of the cardiomyocyte thin layer
  • L(S) represents the objective function with respect to the variable S
  • ⁇ 0 represents the first preset parameter
  • k is a natural number used to control the size of the neighborhood around the voxel position d
  • i is a natural number
  • ⁇ i is the same as the voxel position d +i ⁇ corresponds to the second preset parameter
  • represents a unit voxel
  • represents the norm calculation symbol
  • represents the outer multiplication
  • the inclination angle calculation module is used to calculate the corresponding thin layer inclination angle according to the dual vector expression form of the cardiomyocyte thin layer and the dual vector expression form of the preset reference plane according to the following formula:
  • represents the inclination angle of the thin layer
  • * represents the scalar product operation
  • S r represents the dual vector expression form of the preset reference plane, Indicates the negated or reversed form of S r ;
  • the output display module is used for outputting the dual vector expression forms of all the cardiomyocyte lamellae and the corresponding lamella inclination angles, and obtaining the reconstruction result of the arrangement structure of the cardiomyocyte lamellae of the myocardium.
  • the present invention provides a computer device comprising a communicatively connected memory and a processor, wherein the memory is used to store a computer program, and the processor is used to read the computer program, and execute the first aspect Or any one of the first aspect may design the reconstruction method.
  • the present invention provides a computer-readable storage medium, where instructions are stored on the computer-readable storage medium, and when the instructions are executed on a computer, any one of the first aspect or the first aspect is executed.
  • a possible design of the reconstruction method
  • the present invention provides a computer program product comprising instructions that, when executed on a computer, cause the computer to perform the reconstruction as described in the first aspect or any one of the possible designs of the first aspect method.
  • FIG. 1 is a schematic flowchart of a method for reconstructing a thin layer arrangement structure of cardiomyocytes provided by the present invention.
  • FIG. 2 is a first output example diagram of the reconstruction result of the thin layer arrangement structure of cardiomyocytes provided by the present invention.
  • FIG. 3 is a second output example diagram of the reconstruction result of the myocardial cell thin layer arrangement structure provided by the present invention.
  • FIG. 4 is an output example diagram of a thin layer tilt angle map provided by the present invention.
  • FIG. 5 is an output example diagram of the reconstruction result evaluation diagram provided by the present invention.
  • FIG. 6 is a schematic structural diagram of the reconstruction device of the myocardial cell thin layer arrangement structure provided by the present invention.
  • FIG. 7 is a schematic structural diagram of a computer device provided by the present invention.
  • first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one unit from another. For example, a first element could be referred to as a second element, and similarly a second element could be referred to as a first element, without departing from the scope of example embodiments of this invention.
  • the method for reconstructing the thin layer arrangement structure of cardiomyocytes provided in the first aspect of this embodiment may be, but is not limited to, be executed on a medical image processing device with certain computing resources.
  • the method for reconstructing the arrangement structure of the thin layer of cardiomyocytes may, but is not limited to, include the following steps S101 to S104.
  • the diffusion-weighted image is a medical image collected for a cardiac organ. Specifically, obtaining the geometric algebraic expressions of the diffusion tensors in the diffusion-weighted image, including but not limited to the following steps S1011 to S1013: S1011. Collect the diffusion-weighted image of the myocardium; S1012. Use the nonlinear least squares method from Obtaining a diffusion tensor from the diffusion-weighted image; S1013.
  • the diffusion-weighted image can be acquired by using an existing medical image acquisition device, and the nonlinear least squares method and the feature decomposition method are both existing methods.
  • S represents the double vector expression form of the cardiomyocyte thin layer
  • L(S) represents the objective function with respect to the variable S
  • ⁇ 0 represents the first preset parameter
  • k is a natural number used to control the size of the neighborhood around the voxel position d
  • i is a natural number
  • ⁇ i is the same as the voxel position d +i ⁇ corresponds to the second preset parameter
  • represents a unit voxel
  • represents the norm calculation symbol
  • represents the outer multiplication.
  • step S102 in this embodiment, geometric algebra is used to model the myocardial cell thin layer, that is, G 3 space (the existing concept of space, that is, ⁇ e 1 , e 2 , e 3 ⁇ is set as the three-dimensional space R
  • G 3 space the existing concept of space, that is, ⁇ e 1 , e 2 , e 3 ⁇ is set as the three-dimensional space R
  • three unit double vector bivectors are constructed by extrinsic multiplication: e 1 ⁇ e 2 , e 1 ⁇ e 3 , e 2 ⁇ e 3 , a double vector space ⁇ 2 R 3
  • a unit triple vector e 1 ⁇ e 2 ⁇ e 3 is expanded into a triple vector space ⁇ 3 R 3 .
  • V Take the set ⁇ 1, e 1 , e 2 , e 3 , e 1 ⁇ e 2 , e 1 ⁇ e 3 , e 2 ⁇ e 3 , e 1 ⁇ e 2 ⁇ e 3 ⁇ is used as a basis to form a linear space V with dimension 8.
  • 2 a ⁇ a
  • Geometric products satisfy the distributive and associative laws but not the commutative laws.
  • V constitutes an algebraic structure G 3 , which is called the geometric algebra of the three-dimensional space R 3 , here referred to as the geometric algebra.
  • the double vector bivector as the "directed surface” directly represents the myocardial cell thin sheetlet, Instead of the normal vector in the traditional Euclidean space, or the form of a coordinate point set to represent a face.
  • the benefits brought by this can have the following two points: (1) directly use the double vector bivector to represent the sheetlet of the cardiomyocyte sheetlet in the form of a directed surface, so as to directly calculate the angle (the double vector bivector is the basic calculation unit in geometric algebra, just It directly participates in the operation like a numerical value; the angle between the surfaces can be calculated directly without resorting to vector projection), and avoids the indirect use of the vector on the surface (need to select), or the normal vector of the surface to calculate; (2) can not be collected
  • the slice number of the long-axis of the heart of the image is affected, because the dual vector bivecotr itself has a "direction" without resorting to projection planes or vector references, while for existing methods, when the entire heart is only collected very limited. When data is used, such as 5 slices
  • 2 means it is the standard squared norm or series magnitude of B, and
  • the solution process of the aforementioned optimization problem can be routinely obtained based on the existing algorithm, and finally the dual vector expression form S of the cardiomyocyte thin layer located at the voxel position d is calculated.
  • the thin layers v (1) -v (2) are consistent.
  • represents the inclination angle of the thin layer
  • * represents the scalar product operation
  • S r represents the dual vector expression form of the preset reference plane
  • the preset reference plane preferably adopts the short-axis plane of the heart.
  • the calculation formula of the inclination angle of the thin layer can be replaced by:
  • S and S r are unit duplex vectors respectively, the calculation formula can be further simplified as:
  • the specific output display methods can be, but are not limited to, the following two: (1) first, directly display the dual vector expression form of all myocardial cell thin layers in the form of a circle; then, for the circle at the voxel position d
  • the corresponding preset color is given, as shown in Figure 2 and Figure 3 (in Figure 3, the background layer is the DWI B0 image of the heart, and the The myocardial cell thin layer is from Figure 2 and is selected along the heart wall, so as to observe the change of the inclination angle of the thin layer in the direction across the heart wall) as shown;
  • the quantification calculation process of the layer inclination is conducive to quickly obtaining the reconstruction results of the thin layer arrangement of myocardial cells, which is convenient for real-time reconstruction based on diffusion tensor imaging in clinical medicine, and can be applied to magnetic resonance imaging instruments and diffusion magnetic resonance image analysis/processing software. It is convenient for practical application and promotion.
  • the present embodiment also provides a possible design 1 that can output the reconstruction evaluation result, that is, after calculating the double vector representation of the cardiomyocyte thin layer at voxel position d,
  • the method also includes, but is not limited to, the following steps S201-S204.
  • the numerical magnitude of the standard deviation assign a value corresponding to a preset size (specifically, but not limited to a color depth value) to the thin layer direction similarity and difference index at the voxel position d in the reconstruction result evaluation chart, wherein , the coordinate variable of the reconstruction result evaluation map is the voxel position d, and the thin layer direction similarity index is used to characterize the difference between the cardiomyocyte thin layer at the corresponding voxel position and the cardiomyocyte thin layer at the adjacent voxel position The same degree of direction.
  • a preset size specifically, but not limited to a color depth value
  • the output shows the reconstruction result evaluation graph.
  • a second aspect of this embodiment provides a virtual device for implementing the reconstruction method in the first aspect or any of the possible designs of the first aspect, including an expression acquisition module for sequential communication connection, Dual vector calculation module, tilt angle calculation module and output display module;
  • the expression form obtaining module is used to obtain the geometric algebraic expression form of the diffusion tensor everywhere in the diffusion weighted image, wherein the geometric algebraic expression form includes three eigenvalues ⁇ 1 , ⁇ 2 and ⁇ of the diffusion tensor 3 and the corresponding three unit eigenvectors v (1) , v (2) and v (3) , the three unit eigenvectors v (1) , v (2) and v (3) are perpendicular to each other;
  • the dual vector calculation module is used to calculate the dual vector representation of the cardiomyocyte thin layer at the voxel position d by solving the following optimization problem according to the geometric algebraic representation of the dispersion tensor:
  • S represents the double vector expression form of the cardiomyocyte thin layer
  • L(S) represents the objective function with respect to the variable S
  • ⁇ 0 represents the first preset parameter
  • k is a natural number used to control the size of the neighborhood around the voxel position d
  • i is a natural number
  • ⁇ i is the same as the voxel position d +i ⁇ corresponds to the second preset parameter
  • represents a unit voxel
  • represents the norm calculation symbol
  • represents the outer multiplication
  • the inclination angle calculation module is used to calculate the corresponding thin layer inclination angle according to the dual vector expression form of the cardiomyocyte thin layer and the dual vector expression form of the preset reference plane according to the following formula:
  • represents the inclination angle of the thin layer
  • * represents the scalar product operation
  • S r represents the dual vector expression form of the preset reference plane, Indicates the negated or reversed form of S r ;
  • the output display module is used for outputting the dual vector expression forms of all the cardiomyocyte lamellae and the corresponding lamella inclination angles, and obtaining the reconstruction result of the arrangement structure of the cardiomyocyte lamellae of the myocardium.
  • a third aspect of this embodiment provides a computer device for executing the reconstruction method in the first aspect or any possible design of the first aspect, including a communicatively connected memory and a processor, wherein, The memory is used to store a computer program, and the processor is used to read the computer program and execute the reconstruction method according to the first aspect or any one possible design of the first aspect.
  • the memory may include, but is not limited to, random access memory (Random-Access Memory, RAM), read-only memory (Read-Only Memory, ROM), flash memory (Flash Memory), first-in first-out memory (First Input Memory) First Output, FIFO) and/or first-in-last-out memory (First Input Last Output, FILO), etc.; the processor may not be limited to using a microprocessor of the STM32F105 series.
  • the computer equipment may also include, but is not limited to, a power module, a display screen and other necessary components.
  • a fourth aspect of this embodiment provides a computer-readable storage medium that stores an instruction including the reconstruction method in the first aspect or any one of the possible designs of the first aspect, that is, the computer-readable storage medium is stored thereon
  • the instructions when executed on the computer, perform the reconstruction method as described in the first aspect or any possible design of the first aspect.
  • the computer-readable storage medium refers to a carrier for storing data, which may include, but is not limited to, a floppy disk, an optical disk, a hard disk, a flash memory, a USB flash drive, and/or a memory stick (Memory Stick), etc.
  • the computer may be a general-purpose computer, a special-purpose computer, etc.
  • a fifth aspect of this embodiment provides a computer program product containing instructions, when the instructions are run on a computer, the computer is caused to perform the reconstruction according to the first aspect or any one of the possible designs of the first aspect method.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

Landscapes

  • Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Physics & Mathematics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Geometry (AREA)
  • Computer Graphics (AREA)
  • Radiology & Medical Imaging (AREA)
  • Epidemiology (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Primary Health Care (AREA)
  • Public Health (AREA)
  • Image Analysis (AREA)

Abstract

涉及医学图像处理技术领域,公开了一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质。提供了一种可有效降低薄层建模误差及简化薄层倾角量化计算过程的新型重建方案,即通过基于几何代数G 3空间,使用双重矢量的有向面表示心肌细胞排列形成的薄层,并在此建模方式下重新定义薄层倾角量化计算方式,可以避免带来薄层建模误差,并简化薄层倾角量化计算过程,利于快速得到心肌细胞薄层排列结构的重建结果,方便在临床医学中基于弥散张量成像进行实时重建,进而可应用到磁共振成像仪器及扩散磁共振图像分析/处理软件中,便于实际应用和推广。

Description

一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质 技术领域
本发明属于医学图像处理技术领域,具体涉及一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质。
背景技术
心肌(cardiacmuscle)是由心肌细胞构成的一种肌肉组织,其中,心肌细胞(cardiac muscle cells or cardiomyocytes)是构成心房和心室的最小单位,心脏实现泵血功能得益于心肌细胞在心肌中有序的组织排列。心肌细胞由闰盘(intercalated discs)相连,在心肌中局部呈现方向性,并在心壁中呈现螺旋形态;同时,心肌细胞在心肌中存在另外一种组织形式:它们以约四个心肌细胞厚度为单位堆叠成薄层(sheetlet)。心肌细胞排列成螺旋形态使得心脏在收缩、舒张时心脏会扭转(从心脏短轴short-axis观察);心肌细胞堆叠成薄层结构使得心脏在收缩时相邻薄层相对滑动,趋垂直心壁,心壁变厚,最小化心室容积,以达到最大化向动脉泵血,而在心脏舒张时,相邻薄层相对滑动,逐渐趋向呈现竖立状态,垂直于心脏短轴平面,心壁变薄,最大化心室血容量。前述心肌细胞的两个基本组织形式是心脏实现其功能的微观结构基础。
弥散张量成像(Diffusion Tensor Imaging,DTI)能够以非侵入的方式得到水分子在心肌中的扩散特征,心肌中水分子的扩散特点可以间接地反映心肌细胞的排列结构。从DTI成像输出的弥散加权图像(Diffusion Weighted Image,DWI)出发,可以计算得到弥散张量,DTI成像以弥散张量的形式来表征水在肌肉组织中的扩散。弥散张量可经过特征分解得到三个特征值λ 1、λ 2和λ 3(根据特征值大小降序排列),以及相对应的三个特征向量v (1)、v (2)和v (3)(即第一特征向量primary eigenvector,第二特征向量secondary eigenvector和第三特征向量tertiary eigenvector,均为单位向量,三个特征向量正交),其中,第一特征向量v (1)常被认为和心肌细胞的方向一致,可用于计算螺旋角helix angle(HA),量化心肌细胞排列所呈现的螺旋形态(被认为可以揭示心肌坏死)。而心肌细胞构成的薄层结构动态变化范围大小更能直接体现心脏的泵血能力,即在心脏跳动过程中,薄层相对滑动发生倾斜,倾斜的角度变化范围越大,心脏容积的变化越大,泵血能力越强。
因此有必要基于弥散张量成像对心肌细胞薄层排列结构进行建模、重建和量化,以便用于临床医学,即通过对比机能正常与非正常的心脏,发现心肌结构变化,定量表征心脏机能 是否存在异常及识别机能异常程度等。目前已有薄层建模方式是将特征向量v (1)和v (2)构成的面表示心肌细胞薄层;而已有薄层倾斜角度计算方式为仅用单一的一维特征向量v (2)的投影来量化薄层的角度的变化常称作E2A,E2 angle。
但是上述这些对心肌细胞薄层排列结构进行建模、重建和量化的方法,会存在如下问题:
(1)关于心肌细胞薄层的建模,若用二维面v (1)-v (2)代表心肌细胞薄层,可能会存在一定误差,尤其是在特征值λ 2和λ 3比较接近的情况下(例如,心脏中左右心室连结区域可能出现这样的情况,另外当图像信噪比相对较低的时候,也有可能出现),误差特别明显,这一误差有必要消除或尽量降低;
(2)关于薄层倾斜程度的量化,即首先,薄层倾斜角的定义与计算(即用一维特征向量来表征二维面)是不合适的(即E2A定义并不合理),应该表示面本身、并且利用面本身来定义;其次,计算E2A的过程复杂:需先确定圆周轴circumferential axis,再计算心壁切面(即纵向longitudinal与圆周轴circumferential axis的所在平面)作为特征向量v (1)的投影面,得到投影
Figure PCTCN2021077972-appb-000001
再将投影
Figure PCTCN2021077972-appb-000002
作为法向量,得到另外的心肌细胞薄层交叉面作为特征向量v (2)的投影面,得到投影
Figure PCTCN2021077972-appb-000003
再求这两个投影面的交线,称作“wall-tangent direction”,最后计算投影
Figure PCTCN2021077972-appb-000004
和交线“wall tangent direction”的夹角,得到E2A。
由此在现有心肌细胞薄层排列结构的建模、重建和量化过程中,存在薄层建模有误差及薄层倾角量化计算过程复杂的问题,有必要进行技术改进。
发明内容
为了解决在现有心肌细胞薄层排列结构的建模、重建和量化过程中,存在薄层建模有误差及薄层倾角量化计算过程复杂的问题,本发明目的在于提供一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质,通过基于几何代数G 3空间,使用双重矢量的有向面表示心肌细胞排列形成的薄层,并在此建模方式下重新定义薄层倾角量化计算方式,可以避免带来薄层建模误差,并简化薄层倾角量化计算过程,利于快速得到心肌细胞薄层排列结构的重建结果,方便在临床医学中基于弥散张量成像进行实时重建,进而可应用到磁共振成像仪器及扩散磁共振图像分析/处理软件中,便于实际应用和推广。
第一方面,本发明提供了一种心肌细胞薄层排列结构的重建方法,包括:
获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3), 所述三个单位特征向量v (1)、v (2)和v (3)两两垂直;
根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
Figure PCTCN2021077972-appb-000005
限制条件:||S|| 2=1
式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
Figure PCTCN2021077972-appb-000006
表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
Figure PCTCN2021077972-appb-000007
表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
Figure PCTCN2021077972-appb-000008
表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘;
根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
Figure PCTCN2021077972-appb-000009
式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
Figure PCTCN2021077972-appb-000010
表示S r的求反形式或逆序形式;
输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
基于上述发明内容,提供了一种可有效降低薄层建模误差及简化薄层倾角量化计算过程的新型重建方案,即通过基于几何代数G 3空间,使用双重矢量的有向面表示心肌细胞排列形成的薄层,并在此建模方式下重新定义薄层倾角量化计算方式,可以避免带来薄层建模误差,并简化薄层倾角量化计算过程,利于快速得到心肌细胞薄层排列结构的重建结果,方便在临床医学中基于弥散张量成像进行实时重建,进而可应用到磁共振成像仪器及扩散磁共振图像分析/处理软件中,便于实际应用和推广。
在一个可能的设计中,获取弥散加权图像中各处弥散张量的几何代数表达形式,包括:
采集所述心肌的弥散加权图像;
采用非线性最小二乘法从所述弥散加权图像中获取弥散张量;
对所述弥散加权图像中各处的弥散张量进行特征分解,得到对应的所述三个特征值λ 1、λ 2和λ 3以及所述三个单位特征向量v (1)、v (2)和v (3)
在一个可能的设计中,在计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式之后,所述方法还包括:
针对在所述体素位置d的邻域中各个体素位置d+iδ,分别计算对应的三重矢量级数TM i
Figure PCTCN2021077972-appb-000011
针对所述体素位置d,计算所有三重矢量级数的标准差;
根据所述标准差的数值大小程度,对重建结果评估图中处于体素位置d的薄层方向同异指数赋予对应预设大小的值,其中,所述重建结果评估图的坐标变量为体素位置d,所述薄层方向同异指数用于表征在对应体素位置处心肌细胞薄层与在邻域体素位置处心肌细胞薄层的方向同异程度;
输出展示所述重建结果评估图。
在一个可能的设计中,当所述薄层倾斜角度的规定角度范围介于0~90度之间时,可将所述薄层倾斜角度的计算公式替换为:
Figure PCTCN2021077972-appb-000012
在一个可能的设计中,输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,包括:
以圆面形式直接展示所有心肌细胞薄层的双重矢量表达形式;
针对处于体素位置d的圆面,根据所述心肌细胞薄层的薄层倾斜角度的数值大小程度,赋予对应预设颜色。
在一个可能的设计中,输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,包括:
根据所述薄层倾斜角度的数值大小程度,对薄层倾斜角度图中处于体素位置d的倾角大小指数赋予对应预设大小的值,其中,所述薄层倾斜角度图的坐标变量为体素位置d,所述倾角大小指数用于表征在对应体素位置处心肌细胞薄层与所述预设参考面的夹角大小程度;
输出展示所述薄层倾斜角度图。
在一个可能的设计中,所述预设参考面采用心脏短轴面。
第二方面,本发明提供了一种心肌细胞薄层排列结构的重建装置,包括有依次通信连接的表达形式获取模块、双重矢量计算模块、倾斜角度计算模块和输出展示模块;
所述表达形式获取模块,用于获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3),所述三个单位特征向量v (1)、v (2)和v (3)两两垂直;
所述双重矢量计算模块,用于根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
Figure PCTCN2021077972-appb-000013
限制条件:||S|| 2=1
式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
Figure PCTCN2021077972-appb-000014
表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
Figure PCTCN2021077972-appb-000015
表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
Figure PCTCN2021077972-appb-000016
表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘;
所述倾斜角度计算模块,用于根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
Figure PCTCN2021077972-appb-000017
式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
Figure PCTCN2021077972-appb-000018
表示S r的求反形式或逆序形式;
所述输出展示模块,用于输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
第三方面,本发明提供了一种计算机设备,包括通信相连的存储器和处理器,其中,所述存储器用于存储计算机程序,所述处理器用于读取所述计算机程序,执行如第一方面或第一方面中任意一种可能设计所述的重建方法。
第四方面,本发明提供了一种计算机可读存储介质,所述计算机可读存储介质上存储有指令,当所述指令在计算机上运行时,执行如上第一方面或第一方面中任意一种可能设计的所述重建方法。
第五方面,本发明提供了一种包含指令的计算机程序产品,当所述指令在计算机上运行时,使所述计算机执行如上第一方面或第一方面中任意一种可能设计的所述重建方法。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本发明的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1是本发明提供的心肌细胞薄层排列结构的重建方法流程示意图。
图2是本发明提供的心肌细胞薄层排列结构重建结果的第一种输出示例图。
图3是本发明提供的心肌细胞薄层排列结构重建结果的第二输出示例图。
图4是本发明提供的薄层倾斜角度图的输出示例图。
图5是本发明提供的重建结果评估图的输出示例图。
图6是本发明提供的心肌细胞薄层排列结构的重建装置结构示意图。
图7是本发明提供的计算机设备的结构示意图。
具体实施方式
下面结合附图及具体实施例来对本发明作进一步阐述。在此需要说明的是,对于这些实施例方式的说明虽然是用于帮助理解本发明,但并不构成对本发明的限定。本文公开的特定结构和功能细节仅用于描述本发明的示例实施例。然而,可用很多备选的形式来体现本发明,并且不应当理解为本发明限制在本文阐述的实施例中。
应当理解,尽管本文可能使用术语第一、第二等等来描述各种单元,但是这些单元不应当受到这些术语的限制。这些术语仅用于区分一个单元和另一个单元。例如可以将第一单元称作第二单元,并且类似地可以将第二单元称作第一单元,同时不脱离本发明的示例实施例的范围。
应当理解,对于本文中可能出现的术语“和/或”,其仅仅是一种描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,单独存在B,同时存在A和B三种情况;对于本文中可能出现的术语“/和”,其是描述另一种关联对象关系,表示可以存在两种关系,例如,A/和B,可以表示:单独存在A,单独存在A和B两种情况;另外,对于本文中可能出现的字符“/”,一般表示前后关联对象是一种“或”关系。
应当理解,在本文中若将单元称作与另一个单元“连接”、“相连”或“耦合”时,它可以与另一个单元直相连接或耦合,或中间单元可以存在。相対地,在本文中若将单元称作与另一个单元“直接相连”或“直接耦合”时,表示不存在中间单元。另外,应当以类似方式来解释用于描述单元之间的关系的其他单词(例如,“在……之间”对“直接在……之间”,“相邻”对“直接相邻”等等)。
应当理解,本文使用的术语仅用于描述特定实施例,并不意在限制本发明的示例实施例。若本文所使用的,单数形式“一”、“一个”以及“该”意在包括复数形式,除非上下文明确指示相反意思。还应当理解,若术语“包括”、“包括了”、“包含”和/或“包含了”在本文中被使用时,指定所声明的特征、数量、步骤、操作、单元和/或组件的存在性,并且不排除 一个或多个其他特征、数量、步骤、操作、单元、组件和/或他们的组合存在性或增加。
应当理解,还应当注意到在一些备选可能设计中,所出现的功能/动作可能与附图出现的顺序不同。例如,取决于所涉及的功能/动作,实际上可以实质上并发地执行,或者有时可以以相反的顺序来执行连续示出的两个图。
应当理解,在下面的描述中提供了特定的细节,以便于对示例实施例的完全理解。然而,本领域普通技术人员应当理解可以在没有这些特定细节的情况下实现示例实施例。例如可以在框图中示出系统,以避免用不必要的细节来使得示例不清楚。在其他实例中,可以不以非必要的细节来示出众所周知的过程、结构和技术,以避免使得示例实施例不清楚。
如图1~3所示,本实施例第一方面提供的所述心肌细胞薄层排列结构的重建方法,可以但不限于适用于在具有一定计算资源的医学影像处理设备上执行。所述心肌细胞薄层排列结构的重建方法,可以但不限于包括有如下步骤S101~S104。
S101.获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3),所述三个单位特征向量v (1)、v (2)和v (3)两两垂直。
在所述步骤S101中,所述弥散加权图像即为针对心脏器官采集的医学影像。具体的,获取弥散加权图像中各处弥散张量的几何代数表达形式,包括但不限于有如下步骤S1011~S1013:S1011.采集所述心肌的弥散加权图像;S1012.采用非线性最小二乘法从所述弥散加权图像中获取弥散张量;S1013.对所述弥散加权图像中各处的弥散张量进行特征分解,得到对应的所述三个特征值λ 1、λ 2和λ 3以及所述三个单位特征向量v (1)、v (2)和v (3)。其中,所述弥散加权图像可使用现有医学影像采集设备采集得到,所述非线性最小二乘法和特征分解方法均为现有方法。
S102.根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
Figure PCTCN2021077972-appb-000019
限制条件:||S|| 2=1
式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
Figure PCTCN2021077972-appb-000020
表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
Figure PCTCN2021077972-appb-000021
表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
Figure PCTCN2021077972-appb-000022
表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘。
在所述步骤S102中,本实施例是利用几何代数对心肌细胞薄层进行建模,即用G 3空间(现有空间概念,即设{e 1,e 2,e 3}为三维空间R 3的自然基底,通过外乘构造三个单位双重矢量bivector:e 1∧e 2,e 1∧e 3,e 2∧e 3,张成双重矢量空间∧ 2R 3,一个单位三重矢量:e 1∧e 2∧e 3张成三重矢量空间∧ 3R 3。以集合{1,e 1,e 2,e 3,e 1∧e 2,e 1∧e 3,e 2∧e 3,e 1∧e 2∧e 3}作为基张成一个维数为8的线性空间V。定义两个向量a与b的几何乘积(geometric product)为ab≡a·b+a∧b,且aa=a 2=||a|| 2=a·a,
Figure PCTCN2021077972-appb-000023
几何乘积满足分配律和结合律但不满足交换律。以几何乘积为代数乘,V构成一个代数结构G 3,称之为三维空间R 3的几何代数,这里简称几何代数)中的双重矢量bivector作为“有向面”直接表示心肌细胞薄层sheetlet,而非传统欧式空间中的法向量,或坐标点集的形式来表示一个面。如此带来的好处可有如下两点:(1)直接用双重矢量bivector以有向面的形式表示心肌细胞薄层sheetlet,从而直接计算角度(双重矢量bivector在几何代数中是基本计算单元,就像数值一样直接参与运算;面间的夹角可以直接计算,不必诉诸向量投影),并避免间接利用面上的向量(需要选择),或者面的法向量来计算;(2)可不受采集图像的心脏长轴(long-axis)的层数(slice number)影响,因为双重矢量bivecotr本身有“方向”,无须借助投影面或者向量参考,而对于现存方法,当整个心脏只采集非常有限的数据时,例如5层(slice),计算心壁切面的误差会非常大,特别是当只针对感兴趣的区域仅采集一层的时候,传统方法无法计算投影面,进而无法计算E2A。
在所述步骤S102中,前述优化问题的几何意义:最小化加权体积(平方)和,即对于几何代数向量空间G 3,双重矢量bivector(后文简写B)可以表示有向面,三重矢量trivector(后文简写T)可以表示有向体;一个T可以由外乘(即wedge product,“∧”)构造,即T=B∧v,v表示某个特征向量。||B|| 2表示是B的标准平方squared norm或级数magnitude,||B||可理解为B的面积;类似地,||T||可以理解为T的体积;如此上述优化问题中
Figure PCTCN2021077972-appb-000024
部分表示尽量使待求解的S与
Figure PCTCN2021077972-appb-000025
垂直,这时能得到最大体积,(注意到原优化问题为minimize,且μ 0前有乘负号);
Figure PCTCN2021077972-appb-000026
部分表示尽量使S和
Figure PCTCN2021077972-appb-000027
平行(即周围心肌细胞的局部排列呈现的薄层和S所代表的面平行),即在数学上S和
Figure PCTCN2021077972-appb-000028
在同一个同质化面homogenous space内,这时能得到最小有向体积。此外,前述优化问题的求解过程可基于现有算法常规得到,最终计算出位于所述体素位置d的心肌细胞薄层的双重矢量表达形式S。
在所述步骤S102中,所述第一预设参数μ 0和所述第二预设参数ω i均可预设举例为1, 参数k用于控制邻域大小,降低重建心肌细胞薄层的误差,当k=0时,前述优化问题会退化(即仅考虑当前体素信息而不考虑周围体素信息),此时求解得到的双重矢量bivector所代表的心肌细胞薄层sheetlet与现有定义的薄层面v (1)-v (2)一致。
S103.根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
Figure PCTCN2021077972-appb-000029
式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
Figure PCTCN2021077972-appb-000030
表示S r的求反形式或逆序形式。
在所述步骤S103中,由于
Figure PCTCN2021077972-appb-000031
表示S r的求反形式或逆序形式,有
Figure PCTCN2021077972-appb-000032
所述预设参考面优选采用心脏短轴面。另外,当所述薄层倾斜角度的规定角度范围介于0~90度之间时(即不考虑正负),可将所述薄层倾斜角度的计算公式替换为:
Figure PCTCN2021077972-appb-000033
此外,当S和S r分别为单位二重矢量时,计算公式可进一步简化为:
Figure PCTCN2021077972-appb-000034
S104.输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
在所述步骤S104中,具体输出展示方式可以但不限于有如下两种:(1)先以圆面形式直接展示所有心肌细胞薄层的双重矢量表达形式;然后针对处于体素位置d的圆面,根据所述心肌细胞薄层的薄层倾斜角度的数值大小程度,赋予对应预设颜色,如图2和图3(在图3中,背景图层为心脏的DWI B0图像,图中的心肌细胞薄层来自图2,并沿心壁选取,以便观察薄层倾斜角度在横穿心壁方向上的变化)所示;(2)先根据所述薄层倾斜角度的数值大小程度,对薄层倾斜角度图中处于体素位置d的倾角大小指数赋予对应预设大小的值(具体可以但不限于为颜色深浅值),其中,所述薄层倾斜角度图的坐标变量为体素位置d,所述倾角大小指数用于表征在对应体素位置处心肌细胞薄层与所述预设参考面的夹角大小程度;然后输出展示所述薄层倾斜角度图,如图4所示。
由此基于前述步骤S101~S104所描述的心肌细胞薄层排列结构的重建方法,提供了一种可有效降低薄层建模误差及简化薄层倾角量化计算过程的新型重建方案,即通过基于几何代数G 3空间,使用双重矢量的有向面表示心肌细胞排列形成的薄层,并在此建模方式下重新定义薄层倾角量化计算方式,可以避免带来薄层建模误差,并简化薄层倾角量化计算过程,利于快速得到心肌细胞薄层排列结构的重建结果,方便在临床医学中基于弥散张量成像进行实时重建,进而可应用到磁共振成像仪器及扩散磁共振图像分析/处理软件中,便于实际应 用和推广。
本实施例在前述第一方面的技术方案基础上,还提供了一种可输出重建评估结果的可能设计一,即在计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式之后,所述方法还包括但不限于有如下步骤S201~S204。
S201.针对在所述体素位置d的邻域中各个体素位置d+iδ,分别计算对应的三重矢量级数TM i
Figure PCTCN2021077972-appb-000035
S202.针对所述体素位置d,计算所有三重矢量级数的标准差。
S203.根据所述标准差的数值大小程度,对重建结果评估图中处于体素位置d的薄层方向同异指数赋予对应预设大小的值(具体可以但不限于为颜色深浅值),其中,所述重建结果评估图的坐标变量为体素位置d,所述薄层方向同异指数用于表征在对应体素位置处心肌细胞薄层与在邻域体素位置处心肌细胞薄层的方向同异程度。
S204.输出展示所述重建结果评估图。
在前述步骤S201~S204中,对于某体素位置d,标准差数值越大,表明心肌细胞在心肌内局部邻域内呈现的方向越不一致,如图5所示,特别是在成像仪器信噪比高的情况下,这表明心肌内细胞的微观组织结构复杂(一般出现在心脏的左右心室交汇处)。此外对于右心室,由于心肌较薄,需要综合要考虑实际邻域内的体素个数。
如图6所示,本实施例第二方面提供了一种实现第一方面或第一方面中任意一种可能设计的所述重建方法的虚拟装置,包括有依次通信连接的表达形式获取模块、双重矢量计算模块、倾斜角度计算模块和输出展示模块;
所述表达形式获取模块,用于获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3),所述三个单位特征向量v (1)、v (2)和v (3)两两垂直;
所述双重矢量计算模块,用于根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
Figure PCTCN2021077972-appb-000036
限制条件:||S|| 2=1
式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
Figure PCTCN2021077972-appb-000037
表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
Figure PCTCN2021077972-appb-000038
表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大 小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
Figure PCTCN2021077972-appb-000039
表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘;
所述倾斜角度计算模块,用于根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
Figure PCTCN2021077972-appb-000040
式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
Figure PCTCN2021077972-appb-000041
表示S r的求反形式或逆序形式;
所述输出展示模块,用于输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
本实施例第二方面提供的前述装置的工作过程、工作细节和技术效果,可以参见第一方面或第一方面中任意一种可能设计所述的重建方法,于此不再赘述。
如图7所示,本实施例第三方面提供了一种执行第一方面或第一方面中任意一种可能设计的所述重建方法的计算机设备,包括通信相连的存储器和处理器,其中,所述存储器用于存储计算机程序,所述处理器用于读取所述计算机程序,执行如第一方面或第一方面中任意一种可能设计所述的重建方法。具体举例的,所述存储器可以但不限于包括随机存取存储器(Random-Access Memory,RAM)、只读存储器(Read-Only Memory,ROM)、闪存(Flash Memory)、先进先出存储器(First Input First Output,FIFO)和/或先进后出存储器(First Input Last Output,FILO)等等;所述处理器可以不限于采用型号为STM32F105系列的微处理器。此外,所述计算机设备还可以但不限于包括有电源模块、显示屏和其它必要的部件。
本实施例第三方面提供的前述计算机设备的工作过程、工作细节和技术效果,可以参见第一方面或第一方面中任意一种可能设计所述的重建方法,于此不再赘述。
本实施例第四方面提供了一种存储包含第一方面或第一方面中任意一种可能设计的所述重建方法的指令的计算机可读存储介质,即所述计算机可读存储介质上存储有指令,当所述指令在计算机上运行时,执行如第一方面或第一方面中任意一种可能设计所述的重建方法。其中,所述计算机可读存储介质是指存储数据的载体,可以但不限于包括软盘、光盘、硬盘、闪存、优盘和/或记忆棒(Memory Stick)等,所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。
本实施例第四方面提供的前述计算机可读存储介质的工作过程、工作细节和技术效果,可以参见第一方面或第一方面中任意一种可能设计所述的重建方法,于此不再赘述。
本实施例第五方面提供了一种包含指令的计算机程序产品,当所述指令在计算机上运行时,使所述计算机执行如第一方面或第一方面中任意一种可能设计所述的重建方法。其中,所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。
以上所描述的实施例仅仅是示意性的,若涉及到作为分离部件说明的单元,其可以是或者也可以不是物理上分开的;若涉及到作为单元显示的部件,其可以是或者也可以不是物理单元,即可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。本领域普通技术人员在不付出创造性的劳动的情况下,即可以理解并实施。
以上实施例仅用以说明本发明的技术方案,而非对其限制;尽管参照前述实施例对本发明进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换。而这些修改或者替换,并不使相应技术方案的本质脱离本发明实施例技术方案的精神和范围。
最后应说明的是,本发明不局限于上述可选的实施方式,任何人在本发明的启示下都可得出其他各种形式的产品。上述具体实施方式不应理解成对本发明的保护范围的限制,本发明的保护范围应当以权利要求书中界定的为准,并且说明书可以用于解释权利要求书。

Claims (10)

  1. 一种心肌细胞薄层排列结构的重建方法,其特征在于,包括:
    获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3),所述三个单位特征向量v (1)、v (2)和v (3)两两垂直;
    根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
    Figure PCTCN2021077972-appb-100001
    限制条件:||S|| 2=1
    式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
    Figure PCTCN2021077972-appb-100002
    表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
    Figure PCTCN2021077972-appb-100003
    表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
    Figure PCTCN2021077972-appb-100004
    表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘;
    根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
    Figure PCTCN2021077972-appb-100005
    式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
    Figure PCTCN2021077972-appb-100006
    表示S r的求反形式或逆序形式;
    输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
  2. 如权利要求1所述的重建方法,其特征在于,获取弥散加权图像中各处弥散张量的几何代数表达形式,包括:
    采集所述心肌的弥散加权图像;
    采用非线性最小二乘法从所述弥散加权图像中获取弥散张量;
    对所述弥散加权图像中各处的弥散张量进行特征分解,得到对应的所述三个特征值λ 1、λ 2和λ 3以及所述三个单位特征向量v (1)、v (2)和v (3)
  3. 如权利要求1所述的重建方法,其特征在于,在计算得到位于体素位置d的心肌细 胞薄层的双重矢量表达形式之后,所述方法还包括:
    针对在所述体素位置d的邻域中各个体素位置d+iδ,分别计算对应的三重矢量级数TM i
    Figure PCTCN2021077972-appb-100007
    针对所述体素位置d,计算所有三重矢量级数的标准差;
    根据所述标准差的数值大小程度,对重建结果评估图中处于体素位置d的薄层方向同异指数赋予对应预设大小的值,其中,所述重建结果评估图的坐标变量为体素位置d,所述薄层方向同异指数用于表征在对应体素位置处心肌细胞薄层与在邻域体素位置处心肌细胞薄层的方向同异程度;
    输出展示所述重建结果评估图。
  4. 如权利要求1所述的重建方法,其特征在于,当所述薄层倾斜角度的规定角度范围介于0~90度之间时,可将所述薄层倾斜角度的计算公式替换为:
    Figure PCTCN2021077972-appb-100008
  5. 如权利要求1所述的重建方法,其特征在于,输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,包括:
    以圆面形式直接展示所有心肌细胞薄层的双重矢量表达形式;
    针对处于体素位置d的圆面,根据所述心肌细胞薄层的薄层倾斜角度的数值大小程度,赋予对应预设颜色。
  6. 如权利要求1所述的重建方法,其特征在于,输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,包括:
    根据所述薄层倾斜角度的数值大小程度,对薄层倾斜角度图中处于体素位置d的倾角大小指数赋予对应预设大小的值,其中,所述薄层倾斜角度图的坐标变量为体素位置d,所述倾角大小指数用于表征在对应体素位置处心肌细胞薄层与所述预设参考面的夹角大小程度;
    输出展示所述薄层倾斜角度图。
  7. 如权利要求1所述的重建方法,其特征在于,所述预设参考面采用心脏短轴面。
  8. 一种心肌细胞薄层排列结构的重建装置,其特征在于,包括有依次通信连接的表达形式获取模块、双重矢量计算模块、倾斜角度计算模块和输出展示模块;
    所述表达形式获取模块,用于获取弥散加权图像中各处弥散张量的几何代数表达形式,其中,所述几何代数表达形式包含有弥散张量的三个特征值λ 1、λ 2和λ 3以及相对应的三个单位特征向量v (1)、v (2)和v (3),所述三个单位特征向量v (1)、v (2)和v (3)两两垂直;
    所述双重矢量计算模块,用于根据所述各处弥散张量的几何代数表达形式,通过求解如下优化问题计算得到位于体素位置d的心肌细胞薄层的双重矢量表达形式:
    Figure PCTCN2021077972-appb-100009
    限制条件:||S|| 2=1
    式中,S表示所述心肌细胞薄层的双重矢量表达形式,L(S)表示关于变量S的目标函数,
    Figure PCTCN2021077972-appb-100010
    表示对所述目标函数L(S)求最小值,μ 0表示第一预设参数,
    Figure PCTCN2021077972-appb-100011
    表示在所述体素位置d处弥散张量的第三个单位特征向量,k表示用于控制所述体素位置d周围邻域大小的自然数,i表示自然数,ω i表示与体素位置d+iδ对应的第二预设参数,δ表示单位体素,
    Figure PCTCN2021077972-appb-100012
    表示在体素位置d+iδ处弥散张量的第一个单位特征向量,|| ||表示范数计算符号,∧表示外乘;
    所述倾斜角度计算模块,用于根据所述心肌细胞薄层的双重矢量表达形式和预设参考面的双重矢量表达形式,按照如下公式计算得到对应的薄层倾斜角度:
    Figure PCTCN2021077972-appb-100013
    式中,φ表示所述薄层倾斜角度,*表示标量积运算,S r表示所述预设参考面的双重矢量表达形式,
    Figure PCTCN2021077972-appb-100014
    表示S r的求反形式或逆序形式;
    所述输出展示模块,用于输出展示所有心肌细胞薄层的双重矢量表达形式及对应的薄层倾斜角度,得到所述心肌的心肌细胞薄层排列结构的重建结果。
  9. 一种计算机设备,其特征在于,包括通信相连的存储器和处理器,其中,所述存储器用于存储计算机程序,所述处理器用于读取所述计算机程序,执行如权利要求1~7中任意一项所述的重建方法。
  10. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质上存储有指令,当所述指令在计算机上运行时,执行如权利要求1~7中任意一项所述的重建方法。
PCT/CN2021/077972 2021-02-05 2021-02-25 一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质 WO2022165882A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110163905.0 2021-02-05
CN202110163905.0A CN112819948B (zh) 2021-02-05 2021-02-05 一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质

Publications (1)

Publication Number Publication Date
WO2022165882A1 true WO2022165882A1 (zh) 2022-08-11

Family

ID=75861952

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/077972 WO2022165882A1 (zh) 2021-02-05 2021-02-25 一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN112819948B (zh)
WO (1) WO2022165882A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101666865A (zh) * 2009-09-09 2010-03-10 西北工业大学 一种局部快速行进模式的弥散张量核磁共振图像配准方法
CN103761750A (zh) * 2014-02-14 2014-04-30 华中科技大学 一种心肌质点运动图像与心肌纤维走向图像配准方法
CN104282021A (zh) * 2014-09-28 2015-01-14 深圳先进技术研究院 磁共振弥散张量成像的参数误差估计方法和装置
US9678189B2 (en) * 2013-04-03 2017-06-13 The General Hospital Corporation Mapping cardiac tissue architecture systems and methods

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2914304B1 (fr) * 2007-03-28 2012-11-16 Guerbet Sa Composes pour le diagnostic de maladies liees a l'expression de vcam.
US9576107B2 (en) * 2013-07-09 2017-02-21 Biosense Webster (Israel) Ltd. Model based reconstruction of the heart from sparse samples
US9569887B2 (en) * 2014-05-15 2017-02-14 The Royal Institution For The Advancement Of Learning / Mcgill University Methods of modelling and characterising heart fiber geometry
US10197657B2 (en) * 2015-08-12 2019-02-05 The Regents Of The University Of California Methods and systems for generating a conductivity map of an in vivo tissue

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101666865A (zh) * 2009-09-09 2010-03-10 西北工业大学 一种局部快速行进模式的弥散张量核磁共振图像配准方法
US9678189B2 (en) * 2013-04-03 2017-06-13 The General Hospital Corporation Mapping cardiac tissue architecture systems and methods
CN103761750A (zh) * 2014-02-14 2014-04-30 华中科技大学 一种心肌质点运动图像与心肌纤维走向图像配准方法
CN104282021A (zh) * 2014-09-28 2015-01-14 深圳先进技术研究院 磁共振弥散张量成像的参数误差估计方法和装置

Also Published As

Publication number Publication date
CN112819948A (zh) 2021-05-18
CN112819948B (zh) 2022-08-26

Similar Documents

Publication Publication Date Title
Hoffman et al. VIDA: An environment for multidimensional image display and analysis
Guo et al. Improving cardiac MRI convolutional neural network segmentation on small training datasets and dataset shift: A continuous kernel cut approach
US20050237328A1 (en) Method and system for mesh-to-image registration using raycasting
He et al. Automatic segmentation and quantification of epicardial adipose tissue from coronary computed tomography angiography
JP4856181B2 (ja) 画像データセットからのビューのレンダリング
Frindel et al. Comparison of regularization methods for human cardiac diffusion tensor MRI
Guo et al. Cardiac MRI segmentation with sparse annotations: ensembling deep learning uncertainty and shape priors
WO2022165882A1 (zh) 一种心肌细胞薄层排列结构的重建方法、装置、计算机设备及计算机可读存储介质
Huang et al. Adoption of snake variable model-based method in segmentation and quantitative calculation of cardiac ultrasound medical images
Zhang et al. A visualization system for interactive exploration of the cardiac anatomy
Upendra et al. Automated segmentation of cardiac chambers from cine cardiac MRI using an adversarial network architecture
US9569887B2 (en) Methods of modelling and characterising heart fiber geometry
EP2332123B1 (en) Anatomy-defined automated cpr generation
Wang et al. An ensemble of U-Net architecture variants for left atrial segmentation
Wong et al. Multiview and multiclass image segmentation using deep learning in fetal echocardiography
Adeshina et al. Multimodal 3-D reconstruction of human anatomical structures using surlens visualization system
Bock et al. Coherency-based curve compression for high-order finite element model visualization
Wang et al. Spectrum-sine interpolation framework for DTI processing
Villard et al. Cardiac mesh reconstruction from sparse, heterogeneous contours
Huang et al. Functional analysis of cardiac MR images using SPHARM modeling
Wünsche et al. The visualization of myocardial strain for the improved analysis of cardiac mechanics
Wang Three-dimensional medical CT image reconstruction
Clarysse et al. 3D representation and deformation analysis of the heart walls from X-ray and MR images
Samari Modeling sheets and fibers in biological tissues
Seo et al. Computing diffeomorphic paths with applications to cardiac motion analysis

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: 21923892

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: 21923892

Country of ref document: EP

Kind code of ref document: A1