WO2016106990A1 - Ct成像方法和系统 - Google Patents

Ct成像方法和系统 Download PDF

Info

Publication number
WO2016106990A1
WO2016106990A1 PCT/CN2015/075669 CN2015075669W WO2016106990A1 WO 2016106990 A1 WO2016106990 A1 WO 2016106990A1 CN 2015075669 W CN2015075669 W CN 2015075669W WO 2016106990 A1 WO2016106990 A1 WO 2016106990A1
Authority
WO
WIPO (PCT)
Prior art keywords
image
interest
region
reconstruction
subsystem
Prior art date
Legal status (The legal status 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 status listed.)
Ceased
Application number
PCT/CN2015/075669
Other languages
English (en)
French (fr)
Inventor
张蕴婉
梁栋
胡战利
郑海荣
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shenzhen Institute of Advanced Technology of CAS
Original Assignee
Shenzhen Institute of Advanced Technology of CAS
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 Shenzhen Institute of Advanced Technology of CAS filed Critical Shenzhen Institute of Advanced Technology of CAS
Publication of WO2016106990A1 publication Critical patent/WO2016106990A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T12/00—Tomographic reconstruction from projections
    • G06T12/20—Inverse problem, i.e. transformations from projection space into object space
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2211/00—Image generation
    • G06T2211/40—Computed tomography
    • G06T2211/424—Iterative
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2211/00—Image generation
    • G06T2211/40—Computed tomography
    • G06T2211/444—Low dose acquisition or reduction of radiation dose

Definitions

  • the present invention relates to a CT imaging method and system, and more particularly to an imaging method and imaging system that alternately perform image reconstruction and image restoration using alternating iterations.
  • the CS-based iterative reconstruction algorithm is divided into two types: Algebraic Reconstruction Technique (ART) and Statistical Iterative Reconstruction (SIR).
  • ART Algebraic Reconstruction Technique
  • SIR Statistical Iterative Reconstruction
  • the core idea of the ART reconstruction algorithm is to regard the quadratic projection in the tomographic measurement as a projection set of unknown target functions through the cross section (from the literature: Gordon R, Bender R and Herman G T. Algebraic Reconstruction Techniques (Art) For 3-Dimensional Electron Microscopy and X-Ray Photography [J]. Journal of theoretical biology, 1970, 29: 471-81.).
  • FIG. 1a is an image reconstructed by the ART-TV method for low-dose projection data
  • FIG. 1b and FIG. 1c are partial enlarged views of the reconstruction.
  • the most significant disadvantage of the ART algorithm is that it does not statistically model the data noise, and the reconstruction result is greatly affected by noise. In practical applications, especially in the case of low dose scanning, the data contains a lot of noise.
  • the SIR method is usually applied to the physical effects of the system model and the projection data and noise.
  • the statistical Poisson property of the sound establishes the data fidelity term, and the introduction of the prior information through the regularization constraint item can improve the pathological reconstruction problem to a certain extent, and has better performance in reducing the image noise.
  • the most typical SIR model is the Penalized Weighted Least-Square (PWLS) model.
  • the PWLS model is constructed based on the independence of the projection channel projection data and the maximum a posteriori criterion (from the literature: Wang J, Li T, Lu H, et al. Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography [J]. IEEE Trans Med Imaging, 2006, 25(10): 1272-83.).
  • FIG. 2a shows an image reconstructed by the PWLS-TV method for low-dose projection data
  • FIG. 2b and FIG. 2c are partial enlarged views of the reconstruction.
  • a CT imaging method comprising:
  • Image reconstruction and image restoration are alternately performed according to the image reconstruction model and the image restoration model to obtain a reconstructed image.
  • a CT imaging system comprising:
  • a scanning subsystem for scanning a target object, acquiring all ray projections through the target object, and reconstructing system parameters
  • An image reconstruction subsystem for performing image reconstruction according to an image reconstruction model
  • An iterative subsystem for determining whether the image reconstruction subsystem and the image restoration subsystem alternately perform image reconstruction and image restoration have reached a predetermined number of alternating iterations; and for outputting the reconstructed image.
  • a CT internal region of interest imaging method comprising:
  • the region of interest image reconstruction and the region of interest image restoration are alternately performed to obtain a region of interest reconstructed image.
  • a CT internal region of interest imaging system comprising:
  • a scanning subsystem for scanning an internal region of interest, acquiring all ray projections through the internal region of interest, and reconstructing system parameters
  • the image reconstruction subsystem is configured to perform image reconstruction of the internal region of interest according to the image reconstruction model of the internal region of interest;
  • the image repairing subsystem is configured to perform image repair of the internal region of interest according to the image repairing model of the internal region of interest;
  • An iterative subsystem is configured to determine whether the image reconstruction subsystem and the image restoration subsystem alternately perform image reconstruction of the region of interest and image restoration of the region of interest has reached a predetermined number of alternating iterations; and output the region of interest to reconstruct the image.
  • the CT imaging method and system of the invention scans a target object, acquires all ray projections through the target object, and reconstructs system parameters. By modeling the noise, constructs a CT statistical iterative reconstruction model, and performs detail repair and restoration loss in the iterative process. The boundary information is thus achieved for accurate reconstruction.
  • the CT imaging method and system of the present invention are robust to noisy data and the reconstructed image boundaries are sharper.
  • Figure 1 is an image reconstructed using the ART-TV method for low-dose projection data and an internal region of interest
  • FIG. 3 is a flow chart of a CT imaging method according to an embodiment of the present invention.
  • FIG. 4 is an image reconstructed by a CT imaging method according to an embodiment of the present invention and an internal region of interest;
  • FIG. 5 is a flowchart of step 308 of FIG. 3 according to an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a CT imaging system according to an embodiment of the present invention.
  • first may be referred to as a second client
  • second client may be referred to as a first client, without departing from the scope of the present invention.
  • Both the first client and the second client are clients, but they are not the same client.
  • FIG. 3 it is a flow of a CT imaging method according to an embodiment of the present invention.
  • Step 302 scanning the target object, acquiring all ray projections through the target object and reconstructing system parameters.
  • the scanning method of the CT imaging system may be a parallel beam, a fan beam, or a cone beam; the scanning track may be a circular orbit, a spiral scanning track, or a multi-source static scanning track.
  • the scanning dose can be a normal dose, either a low dose or a sparse angle (Sparse-view).
  • the reconstruction system parameter refers to a geometric parameter required for image reconstruction to construct a CT system matrix.
  • the target object refers to an inner region of interest of the scanned object or the scanned object (ie, the ROI region of the scanned object).
  • the inner region of interest is scanned, all ray projections through the inner region of interest are acquired, and system parameters are reconstructed.
  • all of the ray projections through the target object are: all ray projections that are emitted from the light source, pass through the target object, and reach the detector.
  • an image reconstruction model is constructed. For example, if the target object refers to the internal region of interest, an internal image reconstruction model of the region of interest is constructed, wherein the image reconstruction model for the internal region of interest can be referred to the related description of equation (1) and the image reconstruction model below. If the target object refers to the scanned object, an image reconstruction model of the scanned object is constructed, and an image reconstruction model relating to the scanned object can be referred to the related description of the formula (1) and the image reconstruction model hereinafter.
  • the image reconstruction model is obtained by the following formula (1):
  • y is the acquired ray projection data through the target object
  • x is the currently reconstructed CT image
  • A is the system matrix determined by the acquired system parameters
  • B is a weighting factor containing noise statistics.
  • T represents the transposition operation of the matrix
  • TV(x) represents the total variation function of the variable x
  • ⁇ is the regularization coefficient.
  • the weight value matrix B can be set according to the CT machine projection data noise variance characteristic.
  • TV(x) may be in a general form: Where s and t respectively represent the number of rows and columns of the pixel of the image, wherein ⁇ is a preset constant; it may also be other forms such as high-order TV.
  • noise filtering and artifact removal are simultaneously performed in the optimization solution process, so that the CT imaging method of the present invention is robust to noise, and the image reconstruction in the low dose scanning mode is effective. Better.
  • an image repair model is constructed. If the target object refers to the internal region of interest, then an internal region of interest image repair model is constructed. For the image repair model of the internal region of interest, refer to the relevant descriptions of equation (2) and image repair model below. If the target object refers to the object to be scanned, an image repair model of the scanned object is constructed, and an image repair model relating to the scanned object can be referred to the related description of the formula (2) and the image repair model below.
  • the image repair model is obtained by the following formula (2):
  • ⁇ (x) is the reconstructed image obtained by equation (1)
  • ⁇ (x) is the missing detail information
  • ⁇ (x) is the repaired image.
  • Step 308 performing image reconstruction and image restoration alternately according to the image reconstruction model and the image restoration model to obtain a reconstructed image. For example, if the target object refers to the internal region of interest, then according to the internal region of interest image reconstruction model and the internal region of interest image restoration model, the image reconstruction of the region of interest and the image restoration of the region of interest are alternately performed to obtain the region of interest reconstruction. image. If the target object refers to the scanned object, image reconstruction and image restoration are alternately performed according to the image reconstruction model of the scanned object and the image restoration model of the scanned object to obtain a reconstructed image of the scanned object.
  • the process of "alternating image reconstruction and image restoration to obtain reconstructed images" corresponding to the above two cases can be referred to the following.
  • the performing image reconstruction refers to performing an optimization solution by combining the reconstructed image reconstruction model, specifically referring to an iterative solution of the objective function by using a mathematical optimization method until reaching a preset Termination condition.
  • an iterative method based on gradient descent can be employed.
  • the specific form is as follows (3):
  • ⁇ represents an acceleration factor
  • a TV gradient representing the currently reconstructed CT image x n , where n is a natural number representing the number of iterative operations.
  • the iteration initial value x 0 may be an all-zero or all-one image, or may be a filtered back-projection (FBP) method to reconstruct an image.
  • FBP filtered back-projection
  • the above formula is executed cyclically, and the iterative operation is stopped when the number of loops reaches a preset number of times, and the obtained iterative operation result is taken as the final reconstructed image.
  • iterative solutions can be performed using methods such as conjugate gradient or parabolic substitution.
  • the performing image repair refers to repairing the lost details by combining the constructed image repair model, specifically referring to extracting the detailed information ⁇ (x) lost during the reconstruction process, and adding the image To rebuild the image.
  • ⁇ is the difference image, which may be the difference between the two iterations of the image before and after, or the difference between the images of several iterations
  • ⁇ is the image structure information, and may be extracted from the original image by using an algorithm such as segmentation. Represents a point multiplication operation.
  • the image reconstruction and image restoration are alternately performed, and the number of iterations may be determined according to a specific situation.
  • the reconstruction model can be iterated once and repaired once (where the difference image represents the image difference before and after the reconstruction model is iterated once), and the whole iteration is 50 times; the reconstruction model can also be iterated 10 times and repaired once (where The difference image represents the image difference before and after the reconstruction model is iterated 10 times, and the whole iteration is 20 times.
  • FIG. 4a is an image reconstructed by the CT imaging method of the present invention for low-dose projection data
  • FIG. 4b and FIG. 4c are partial enlarged views.
  • detail restoration is performed in an iterative process to restore the lost boundary information, so that the reconstructed image boundary is clearer.
  • step 308 in FIG. 3 it is the flow of step 308 in FIG. 3 in accordance with an embodiment of the present invention.
  • Step 502 Perform image reconstruction according to the image reconstruction model.
  • the image reconstruction may be performed by using the iterative method described in the foregoing formula (3), and the above formula (3) is executed cyclically, and the iterative operation is stopped when the number of loops reaches a preset number of times, and the obtained iterative operation result is obtained.
  • iterative solution can also be performed using methods such as conjugate gradient or parabolic substitution.
  • Step 504 performing image repair according to the image repair model.
  • the detailed information ⁇ (x) lost during the reconstruction process can be extracted by using the foregoing formula (4), and added to the reconstructed image, and the image repair is performed according to the image repair model of the foregoing formula (2).
  • step 506 it is determined whether the predetermined number of alternating iterations has been reached. If not, step 502 and step 504 are still performed to perform image reconstruction and image restoration alternately.
  • the number of alternate iterations can be determined on a case-by-case basis.
  • the reconstruction model can be iterated once and repaired once (where the difference image represents the image difference before and after the reconstruction model is iterated once), and the whole iteration is 50 times; the reconstruction model can also be iterated 10 times and repaired once (where The difference image represents the image difference before and after the reconstruction model is iterated 10 times, and the whole iteration is 20 times.
  • Step 508 if it is determined in step 506 that the predetermined number of alternating iterations has been reached, a reconstructed image is obtained.
  • the reconstructed image is less noisy and the boundary is sharper.
  • FIG. 6 is a schematic structural diagram of a CT imaging system according to an embodiment of the present invention.
  • the CT imaging system 600 includes a scanning subsystem 602, an image reconstruction subsystem 604, an image restoration subsystem 606, and an iterative subsystem 608.
  • the scanning subsystem 602 is used to scan a target object, acquire all ray projections through the target object, and reconstruct system parameters.
  • the scanning subsystem 602 of the CT imaging system 600 may scan in a parallel beam, a fan beam, or a cone beam; the scanning track may be a circular orbit, a spiral scanning track, or a multi-source static scanning track.
  • the scanning dose can be a normal dose, either a low dose or a sparse angle (Sparse-view).
  • the reconstruction system parameter refers to a geometric parameter required for image reconstruction to construct a CT system matrix.
  • the target object refers to an inner region of interest of the scanned object or the scanned object.
  • the scanning subsystem described above is used to scan the interior region of interest, acquire all ray projections through the interior region of interest, and reconstruct system parameters.
  • the scanning subsystem is used to scan the scanned object, acquire all ray projections by scanning the object, and reconstruct system parameters.
  • all of the ray projections through the target object are: all ray projections that are emitted from the light source, pass through the target object, and reach the detector.
  • the image reconstruction subsystem 604 is configured to perform image reconstruction in accordance with the image reconstruction model. If the target object refers to the inner region of interest, the image reconstruction subsystem is configured to reconstruct the image of the inner region of interest according to the image reconstruction model of the inner region of interest. If the target object refers to the scanned object, the image reconstruction subsystem is configured to perform image reconstruction of the scanned object according to the image reconstruction model of the scanned object.
  • the image reconstruction model is obtained by the above formula (1).
  • the image restoration subsystem 606 is configured to perform image restoration according to the image restoration model. For example, if the target object refers to the inner region of interest, the image repair subsystem is configured to perform image repair of the inner region of interest according to the image repair model of the inner region of interest. If the target object refers to the scanned object, the image repair subsystem is configured to perform image restoration of the scanned object according to the image repair model of the scanned object.
  • the above-mentioned formula (4) may be used to extract the detail information ⁇ (x) lost during the reconstruction process, and add it to the reconstructed image according to the foregoing formula (2).
  • the image restoration model performs image restoration.
  • the iterative subsystem 608 is operative to determine if a predetermined number of alternate iterations have been reached, and if not, the iterative subsystem 608 instructs the image reconstruction subsystem 604 and the image restoration subsystem 606 to alternate image reconstruction and image restoration. If the predetermined number of alternating iterations has been reached, the reconstructed image is output. For example, if the target object refers to an internal region of interest, the iterative subsystem is configured to determine whether the image reconstruction subsystem and the image restoration subsystem alternately perform image reconstruction of the region of interest and image restoration of the region of interest has reached a predetermined alternating iteration. The number of times; and the reconstructed image for outputting the region of interest.
  • the iterative subsystem is configured to determine whether the image reconstruction subsystem and the image restoration subsystem alternately perform scanning object image reconstruction and scan object image restoration has reached a predetermined number of alternating iterations; Outputs a reconstructed image of the scanned object.
  • the number of alternate iterations can be determined on a case-by-case basis.
  • the reconstruction model can be iterated once and repaired once (where the difference image represents the image difference before and after the reconstruction model is iterated once), and the whole iteration is 50 times; the reconstruction model can also be iterated 10 times and repaired once (where The difference image represents the image difference before and after the reconstruction model is iterated 10 times, and the whole iteration is 20 times.
  • the reconstructed image is less noisy and the boundaries are sharper.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Apparatus For Radiation Diagnosis (AREA)

Abstract

一种CT成像方法，包括：扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数；构建图像重建模型；构建图像修复模型；依据图像重建模型和图像修复模型，交替进行图像重建和图像修复，以获得重建图像。本发明还提供一种相应的CT成像系统。

Description

CT成像方法和系统 【技术领域】
本发明涉及一种CT的成像方法和系统，特别地，涉及一种利用交替迭代交替进行图像重建和图像修复的成像方法和成像系统。
【背景技术】
近年来，针对CT图像优质重建的研究主要基于压缩感知(Compress Sensing，CS)算法。研究证明了一个信号如果是稀疏表示的，则可以利用全变分(Total Variation，TV)最小化的方法从少量测量数据中精确重建信号。
基于CS的迭代重建算法，总体上分为代数迭代重建算法(Algebraic Reconstruction Technique，ART)和统计迭代重建算法(Statistic Iterative Reconstruction，SIR)两种。其中，ART重建算法的核心思想是将断层测量中的二次投影视为一些通过横截面的未知目标函数的投影集(出自文献：Gordon R,Bender R and Herman G T.Algebraic Reconstruction Techniques(Art)for 3-Dimensional Electron Microscopy and X-Ray Photography[J].Journal of theoretical biology,1970,29:471-81.)。
如图1所示，其中图1a为低剂量投影数据采用ART-TV方法重建的图像；图1b和图1c为该重建的局部放大图。
ART算法最显著的缺点是没有对数据噪声进行统计建模，重建结果受噪声影响比较大。在实际应用中，尤其是在低剂量扫描情况下，数据含有大量噪声。相较于ART重建算法，通常SIR方法针对系统模型的物理效应和投影数据和噪 声的统计泊松特性建立数据保真项，同时通过正则化约束项引入先验信息，能够在一定程度上改善病态重建问题，在降低图像噪声方面有更佳的表现。最为典型的SIR模型是惩罚加权最小二乘重建模型(Penalized Weighted Least-Square，PWLS)，PWLS模型根据探测器通道投影数据的独立性以及最大后验准则构建(出自文献：Wang J,Li T,Lu H,et al.Penalized weighted least-squares approach to sinogram noise reduction and image reconstruction for low-dose X-ray computed tomography[J].IEEE Trans Med Imaging,2006,25(10):1272-83.)。
如图2所示，其中图2a为低剂量投影数据采用采用PWLS-TV方法重建的图像；图2b和图2c为该重建的局部放大图。
然而，无论是ART算法还是SIR算法，在抑制噪声和伪影的同时，会损失掉图像原有的细节结构，使图像边界变模糊。因此，在抑制噪声伪影的同时保持原有边界信息将成为CT图像优质重建的需要。
【发明内容】
基于此，有必要提供一种简单、精确的CT成像方法和系统。
一种CT成像方法，包括：
扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数；
构建图像重建模型；
构建图像修复模型；
依据图像重建模型和图像修复模型，交替进行图像重建和图像修复，以获得重建图像。
一种CT成像系统，包括：
扫描子系统，用于扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数；
图像重建子系统，用于依据图像重建模型，进行图像重建；
图像修复子系统，用于依据图像修复模型，进行图像修复；
迭代子系统，用于判断所述图像重建子系统和图像修复子系统交替进行图像重建和图像修复是否已经达到预定的交替迭代次数；以及用于输出重建图像。
一种CT内部感兴趣区域成像方法，其包括：
扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数；
构建内部感兴趣区域图像重建模型；
构建内部感兴趣区域图像修复模型；
依据内部感兴趣区域图像重建模型和内部感兴趣区域图像修复模型，交替进行感兴趣区域图像重建和感兴趣区域图像修复，以获得感兴趣区域重建图像。
一种CT内部感兴趣区域成像系统，其包括：
扫描子系统，用于扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数；
图像重建子系统，用于依据内部感兴趣区域图像重建模型，进行内部感兴趣区域图像重建；
图像修复子系统，用于依据内部感兴趣区域图像修复模型，进行内部感兴趣区域图像修复；
迭代子系统，用于判断所述图像重建子系统和图像修复子系统交替进行感兴趣区域图像重建和感兴趣区域图像修复是否已经达到预定的交替迭代次数；以及用于输出感兴趣区域重建图像。
本发明的CT成像方法和系统，扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数，通过对噪声进行建模，构建CT统计迭代重建模型，并在迭代过程进行细节修复，还原损失的边界信息，从而实现精确重建。本发明的CT成像方法和系统对含噪声数据鲁棒，重建图像边界更清晰。
【附图说明】
图1为低剂量投影数据采用ART-TV方法重建得到的图像以及内部感兴趣区域；
图2为低剂量投影数据采用PWLS-TV方法重建得到的图像以及内部感兴趣区域；
图3为本发明一种实施方式的CT成像方法的流程图；
图4为低剂量投影数据采用本发明实施方式的CT成像方法重建得到的图像以及内部感兴趣区域；
图5为本发明一种实施方式的图3中步骤308的流程；
图6为本发明一种实施方式的CT成像系统的结构示意图。
【具体实施方式】
为了使本发明的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本发明进行进一步详细说明。应当理解，此处所描述的具体实施例仅仅用以解释本发明，并不用于限定本发明。
可以理解，本发明所使用的术语“第一”、“第二”等可在本文中用于描述各种元件，但这些元件不受这些术语限制。这些术语仅用于将第一个元件与另一个元件区分。举例来说，在不脱离本发明的范围的情况下，可以将第一客户端称为第二客户端，且类似地，可将第二客户端称为第一客户端。第一客户端和第二客户端两者都是客户端，但其不是同一客户端。
如图3所示，其为本发明一种实施方式的CT成像方法的流程。
步骤302，扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数。
具体地，CT成像系统进行扫描的方式可以是平行束、扇束、锥束；扫描轨道可以是圆轨道、螺旋扫描轨道，也可以是多源静态扫描轨道。扫描剂量可以是正常剂量，可以是低剂量，也可以是稀疏角度(Sparse-view)。
根据本发明的一种实施方式，所述重建系统参数是指进行图像重建所需要的几何参数，用以构建CT系统矩阵。
根据本发明的一种实施方式，所述目标物体是指被扫描物体或者被扫描物体的内部感兴趣区域(即被扫描物体的ROI区域)。例如，上述步骤302中，扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数。
根据本发明的一种实施方式，所述通过目标物体的所有射线投影是指：从光源发出，穿过目标物体，到达探测器的所有射线投影。
步骤304，构建图像重建模型。例如，若目标物体是指内部感兴趣区域，那么构建内部感兴趣区域图像重建模型，其中有关内部感兴趣区域图像重建模型可参见下文中有关式(1)和图像重建模型的相关说明。若目标物体是指被扫描物体，那么构建扫描物体的图像重建模型，其中有关扫描物体的图像重建模型可参见下文中有关式(1)和图像重建模型的相关说明。
具体地，根据本发明的一种实施方式，所述图像重建模型由以下式(1)得到：
Figure PCTCN2015075669-appb-000001
其中，y是所获取的通过目标物体的所有射线投影数据，x是当前重建的CT图像，A是由所获取的系统参数确定的系统矩阵，B为加权因子，包含噪声统计信息。T表示矩阵的转置运算，TV(x)表示变量x的全变分函数，λ为正则化系数。
在可选的其他实施方式中，为构建图像重建模型，权重值矩阵B可以根据CT机投影数据噪声方差特性设定。
在可选的其他实施方式中，为构建图像重建模型，TV(x)可以是一般的形式：
Figure PCTCN2015075669-appb-000002
其中s、t分别表示图像像素点所在的行数和列数，其中的δ为预设的常数；也可以是高阶TV等其他形式。
通过在重建目标函数加入噪声统计信息，在优化求解过程中同时进行噪声滤除和伪影去除，使得本发明的CT成像方法对噪声鲁棒，在处理低剂量扫描模式下的图像重建时，效果更佳。
步骤306，构建图像修复模型。若目标物体是指内部感兴趣区域，那么构建内部感兴趣区域图像修复模型。其中有关内部感兴趣区域图像修复模型可参见下文中有关式(2)和图像修复模型的相关说明。若目标物体是指被扫描物体，那么构建扫描物体的图像修复模型，其中有关扫描物体的图像修复模型可参见下文中有关式(2)和图像修复模型的相关说明。
根据本发明的一种实施方式，所述图像修复模型由以下式(2)得到：
Θ(x)＝Φ(x)+Ψ(x)………………………………(2)
其中，Φ(x)是由式(1)所获得的重建图像，Ψ(x)是丢失的细节信息，Θ(x)是修复后的图像。
步骤308，依据图像重建模型和图像修复模型，交替进行图像重建和图像修复，以获得重建图像。例如，若目标物体是指内部感兴趣区域，那么依据内部感兴趣区域图像重建模型和内部感兴趣区域图像修复模型，交替进行感兴趣区域图像重建和感兴趣区域图像修复，以获得感兴趣区域重建图像。若目标物体是指被扫描物体，那么依据扫描物体的图像重建模型和扫描物体的图像修复模型，交替进行图像重建和图像修复，以获得扫描物体的重建图像。以上两种情况对应的“交替进行图像重建和图像修复，以获得重建图像”过程均可参照下文所述。
根据本发明的一种可选的实施方式，所述进行图像重建，是指结合构建的图像重建模型进行优化求解，具体是指采用数学优化方法对目标函数进行迭代求解，直到到达预先设定的终止条件。
在本发明的一种实施方式中，可以采用基于梯度下降的迭代方法。具体形式为下式(3)：
Figure PCTCN2015075669-appb-000003
其中，η表示加速因子，
Figure PCTCN2015075669-appb-000004
表示当前重建的CT图像xn的TV梯度，n为自然数，表示迭代运算的次数。
其中，迭代初始值x0可以是全0或者全1图像，也可以是滤波反投影(Filtered Back-Projection，FBP)方法重建图像。
循环执行上述公式，当循环次数达到预设的次数时即停止迭代运算，并将所得到的迭代运算结果作为最终的重建图像。
除此之外，还可采用共轭梯度或者抛物替代等方法进行迭代求解。
根据本发明的一种可选的实施方式，所述进行图像修复，是指结合构建的图像修复模型修复丢失的细节，具体是指提取重建过程中丢失的细节信息Ψ(x)，将其加入到重建图像中。
其中Ψ(x)的提取算法可以采用以下式(4)：
Figure PCTCN2015075669-appb-000005
其中ν是差值图像，可以是前后两次迭代图像之差，也可以是若干次迭代图像之差，μ是图像结构信息，可采用分割等算法从原图像中提取，
Figure PCTCN2015075669-appb-000006
表示点乘运算。
在本发明的一种实施方式中，所述交替进行图像重建和图像修复，其迭代次数可以根据具体情况而定。例如，可以使重建模型迭代1次，修复1次(其中差值图像表示重建模型迭代1次前后的图像差值)，整体迭代50次；也可以使重建模型迭代10次，修复1次(其中差值图像表示重建模型迭代10次前后的图像差值)，整体迭代20次。
如图4所示，其中图4a为低剂量投影数据采用本发明的CT成像方法重建的图像；图4b和图4c为局部放大图。如图4b、图4c中的箭头所示可见，根据本发明的CT成像方法，在迭代过程进行细节修复，还原损失的边界信息，使得重建图像边界更清晰。
如图5所示，其为根据本发明一种实施方式的图3中步骤308的流程。
步骤502，依据图像重建模型，进行图像重建。
具体地，可以采用前述的式(3)所述的迭代方式进行图像重建，循环执行上述公式(3)，当循环次数达到预设的次数时即停止迭代运算，并将所得到的迭代运算结果作为最终的重建图像。根据本发明的其他实施方式，还可采用共轭梯度或者抛物替代等方法进行迭代求解。
步骤504，依据图像修复模型，进行图像修复。
具体地，可以利用前述式(4)来提取重建过程中丢失的细节信息Ψ(x)，并将其加入到重建图像中，依据前述的式(2)的图像修复模型进行图像修复。
步骤506，判断是否已经达到预定的交替迭代次数，若未达到，则仍需进行步骤502、步骤504，交替进行图像重建和图像修复。
交替迭代次数可以根据具体情况而定。例如，可以使重建模型迭代1次，修复1次(其中差值图像表示重建模型迭代1次前后的图像差值)，整体迭代50次；也可以使重建模型迭代10次，修复1次(其中差值图像表示重建模型迭代10次前后的图像差值)，整体迭代20次。
步骤508，若步骤506中判断已经达到预定的交替迭代次数，则获得重建图像。
通过本发明的该CT成像方法，重建图像噪声更少，边界更加清晰。
如图6所示，其为本发明一种实施方式的CT成像系统的结构示意图。
该CT成像系统600包括扫描子系统602、图像重建子系统604、图像修复子系统606，以及迭代子系统608。
扫描子系统602用于扫描目标物体，获取通过目标物体的所有射线投影以及重建系统参数。
具体地，CT成像系统600的扫描子系统602进行扫描的方式可以是平行束、扇束、锥束；扫描轨道可以是圆轨道、螺旋扫描轨道，也可以是多源静态扫描轨道。扫描剂量可以是正常剂量，可以是低剂量，也可以是稀疏角度(Sparse-view)。
根据本发明的一种实施方式，所述重建系统参数是指进行图像重建所需要的几何参数，用以构建CT系统矩阵。
根据本发明的一种实施方式，所述目标物体是指被扫描物体或者被扫描物体的内部感兴趣区域。例如，若目标物体是指内部感兴趣区域，那么上述扫描子系统用于扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数。若目标物体是指被扫描物体，那么上述扫描子系统用于扫描扫描物体，获取通过扫描物体的所有射线投影以及重建系统参数。
根据本发明的一种实施方式，所述通过目标物体的所有射线投影是指：从光源发出，穿过目标物体，到达探测器的所有射线投影。
图像重建子系统604用于依据图像重建模型，进行图像重建。若目标物体是指内部感兴趣区域，那么图像重建子系统用于依据内部感兴趣区域图像重建模型，进行内部感兴趣区域图像重建。若目标物体是指被扫描物体，那么图像重建子系统用于依据扫描物体的图像重建模型，进行扫描物体的图像重建。
具体地，根据本发明的一种实施方式，所述图像重建模型由前述的式(1)得到。
通过在重建目标函数加入噪声统计信息，在优化求解过程中同时进行噪声 滤除和伪影去除，使得本发明的CT成像方法对噪声鲁棒，在处理低剂量扫描模式下的重建时，效果更佳。
图像修复子系统606用于依据图像修复模型，进行图像修复。例如若目标物体是指内部感兴趣区域，那么图像修复子系统用于依据内部感兴趣区域图像修复模型，进行内部感兴趣区域图像修复。若目标物体是指被扫描物体，那么图像修复子系统用于依据扫描物体的图像修复模型，进行扫描物体的图像修复。
根据本发明的一种实施方式，具体地，可以利用前述式(4)来提取重建过程中丢失的细节信息Ψ(x)，并将其加入到重建图像中，依据前述的式(2)的图像修复模型进行图像修复。
迭代子系统608用于判断是否已经达到预定的交替迭代次数，若未达到，则迭代子系统608指示图像重建子系统604和图像修复子系统606交替进行图像重建和图像修复。若已经达到预定的交替迭代次数，则输出重建图像。例如，若目标物体是指内部感兴趣区域，那么迭代子系统用于判断所述图像重建子系统和图像修复子系统交替进行感兴趣区域图像重建和感兴趣区域图像修复是否已经达到预定的交替迭代次数；以及用于输出感兴趣区域重建图像。若目标物体是指被扫描物体，那么迭代子系统用于判断所述图像重建子系统和图像修复子系统交替进行扫描物体图像重建和扫描物体图像修复是否已经达到预定的交替迭代次数；以及用于输出扫描物体的重建图像。
交替迭代次数可以根据具体情况而定。例如，可以使重建模型迭代1次，修复1次(其中差值图像表示重建模型迭代1次前后的图像差值)，整体迭代50次；也可以使重建模型迭代10次，修复1次(其中差值图像表示重建模型迭代10次前后的图像差值)，整体迭代20次。
通过本发明的该CT成像系统，重建图像噪声更少，边界更加清晰。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序控制相关的硬件来完成的，所述的程序可存储于一计算机可读取存储介质中，该程序在执行时，可包括如上述各方法的实施例的流程。其中，所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory，ROM)或随机存储记忆体(Random Access Memory，RAM)等。
以上所述实施例仅表达了本发明的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对本发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本发明构思的前提下，还可以做出若干变形和改进，这些都属于本发明的保护范围。因此，本发明专利的保护范围应以所附权利要求为准。
Figure PCTCN2015075669-appb-000007

Claims (35)

  1. 其中ν是差值图像，μ是图像结构信息，
    Figure PCTCN2015075669-appb-100001
    表示点乘运算。
  2. 如权利要求7所述的CT成像方法，其特征在于：所述差值图像ν为前后两次迭代图像之差或多次迭代图像之差。
  3. 如权利要求7所述的CT成像方法，其特征在于：所述图像结构信息μ采用分割等算法从原图像中提取。
  4. 如权利要求1所述的CT成像方法，其特征在于：所述交替进行图像重建和图像修复采用基于梯度下降的迭代方法：
    xn＝xn-1-η×(AT(B(y-Axn-1)))-λ×▽TV(xn-1)；
    其中，η为加速因子，▽TV(xn)为当前重建的CT图像xn的TV梯度，n为自然数，表示迭代运算的次数。
  5. 如权利要求10所述的CT成像方法，其特征在于：迭代初始值x0为以下各项中的任一项：全0图像，全1图像，滤波反投影方法重建的图像。
  6. 如权利要求1所述的CT成像方法，其特征在于：所述交替进行图像重建和图像修复采用共轭梯度或者抛物替代的迭代方法。
  7. 如权利要求1所述的CT成像方法，其特征在于：所述交替进行图像重建和图像修复包括：
    循环执行所述依据图像重建模型进行图像重建和所述依据图像修复模型进行图像修复；
    当循环次数达到预设的次数时，停止迭代运算，并将所得到的迭代运算结果作为所获得的重建图像。
  8. 一种CT成像系统，其特征在于，包括：
    扫描子系统，用于扫描目标物体，获取通过该物体的所有射线投影以及重建系统参数；
    图像重建子系统，用于依据图像重建模型，进行图像重建；
    图像修复子系统，用于依据图像修复模型，进行图像修复；
    迭代子系统，用于判断所述图像重建子系统和图像修复子系统交替进行图像重建和图像修复是否已经达到预定的交替迭代次数；以及用于输出重建图像。
  9. 如权利要求14所述的CT成像系统，其特征在于：若迭代子系统判断 所述图像重建子系统和图像修复子系统交替进行图像重建和图像修复未达到预定的交替迭代次数，则所述迭代子系统指示所述图像重建子系统和图像修复子系统交替进行图像重建和图像修复。
  10. 如权利要求14所述的CT成像系统，其特征在于：若迭代子系统判断所述图像重建子系统和图像修复子系统交替进行图像重建和图像修复已经达到预定的交替迭代次数，则输出所述重建图像。
  11. 如权利要求14所述的CT成像系统，其特征在于：所述通过目标物体的所有射线投影是：从光源发出，穿过目标物体，到达探测器的所有射线投影。
  12. 如权利要求14所述的CT成像系统，其特征在于：所述图像重建模型为：
    Figure PCTCN2015075669-appb-100002
    其中，y是获取的通过目标物体的所有射线投影数据，x是当前重建的CT图像，A是由获取的系统参数确定的系统矩阵，B为加权因子，包含噪声统计信息。T表示矩阵的转置运算，TV(x)表示变量x的全变分函数，λ为正则化系数。
  13. 如权利要求18所述的CT成像系统，其特征在于：所述变量x的全变分函数TV(x)为：
    Figure PCTCN2015075669-appb-100003
    其中s、t分别表示图像像素点所在的行数和列数，其中的δ为预设的常数。
  14. 如权利要求14所述的CT成像系统，其特征在于：图像修复模型为：
    Θ(x)＝Φ(x)+Ψ(x)；
    其中，Φ(x)所述图像重建模型，Ψ(x)是丢失的细节信息，Θ(x)是修复后的图像。
  15. 如权利要求14所述的CT成像系统，其特征在于：所述进行图像修复，包括提取重建过程中丢失的细节信息Ψ(x)，将其加入到重建图像中。
  16. 如权利要求21所述的CT成像系统，其特征在于：所述丢失的细节信息Ψ(x)的提取算法为：
    Figure PCTCN2015075669-appb-100004
    其中ν是差值图像，μ是图像结构信息，
    Figure PCTCN2015075669-appb-100005
    表示点乘运算。
  17. 如权利要求21所述的CT成像系统，其特征在于：所述差值图像ν为前后两次迭代图像之差或多次迭代图像之差。
  18. 如权利要求21所述的CT成像系统，其特征在于：所述图像结构信息μ采用分割等算法从原图像中提取。
  19. 如权利要求14所述的CT成像系统，其特征在于：所述交替进行图像重建和图像修复采用基于梯度下降的迭代方法：
    xn＝xn-1-η×(AT(B(y-Axn-1)))-λ×▽TV(xn-1)；
    其中，η为加速因子，▽TV(xn)为当前重建的CT图像xn的TV梯度，n为自然数，表示迭代运算的次数。
  20. 如权利要求24所述的CT成像系统，其特征在于：迭代初始值x0为以下各项中的任一项：全0图像，全1图像，滤波反投影方法重建的图像。
  21. 如权利要求14所述的CT成像系统，其特征在于：所述交替进行图像重建和图像修复采用共轭梯度或者抛物替代的迭代方法。
  22. 一种CT内部感兴趣区域成像方法，其特征在于，包括：
    扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数；
    构建内部感兴趣区域图像重建模型；
    构建内部感兴趣区域图像修复模型；
    依据内部感兴趣区域图像重建模型和内部感兴趣区域图像修复模型，交替进行感兴趣区域图像重建和感兴趣区域图像修复，以获得感兴趣区域重建图像。
  23. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述通过内部感兴趣区域的所有射线投影为：从光源发出，穿过内部感兴趣区域，到达探测器的所有射线投影。
  24. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述内部感兴趣区域图像重建模型为：
    Figure PCTCN2015075669-appb-100006
    其中，y是获取的通过感兴趣区域的所有射线投影数据，x是当前重建的CT 图像，A是由获取的系统参数确定的系统矩阵，B为加权因子，包含噪声统计信息。T表示矩阵的转置运算，TV(x)表示变量x的全变分函数，λ为正则化系数。
  25. 如权利要求30所述的CT内部感兴趣区域成像方法，其特征在于：所述变量x的全变分函数TV(x)为：
    Figure PCTCN2015075669-appb-100007
    其中s、t分别表示图像像素点所在的行数和列数，其中的δ为大于0小于10-8的常数。
  26. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：内部感兴趣区域图像修复模型为：
    Θ(x)＝Φ(x)+Ψ(x)；
    其中，Φ(x)所述内部感兴趣区域图像重建模型，Ψ(x)是丢失的细节信息，Θ(x)是修复后的图像。
  27. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述进行感兴趣区域图像修复，包括提取重建过程中丢失的细节信息Ψ(x)，将其加入到重建图像中。
  28. 如权利要求33所述的CT内部感兴趣区域成像方法，其特征在于：所述丢失的细节信息Ψ(x)的提取算法为：
    Figure PCTCN2015075669-appb-100008
    其中ν是差值图像，μ是图像结构信息，
    Figure PCTCN2015075669-appb-100009
    表示点乘运算。
  29. 如权利要求34所述的CT内部感兴趣区域成像方法，其特征在于：所述差值图像ν为前后两次迭代图像之差或多次迭代图像之差。
  30. 如权利要求34所述的CT内部感兴趣区域成像方法，其特征在于：所述图像结构信息μ采用分割等算法从原图像中提取。
  31. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述交替进行感兴趣区域图像重建和感兴趣区域图像修复采用基于梯度下降的迭代方法：
    xn＝xn-1-η×(AT(B(y-Axn-1)))-λ×▽TV(xn-1)；
    其中，η为加速因子，▽TV(xn)为当前重建的CT图像xn的TV梯度，n为自然数，表示迭代运算的次数。
  32. 如权利要求37所述的CT内部感兴趣区域成像方法，其特征在于：迭代初始值x0为以下各项中的任一项：全0图像，全1图像，滤波反投影方法重建的图像。
  33. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述交替进行感兴趣区域图像重建和感兴趣区域图像修复采用共轭梯度或者抛物替代的迭代方法。
  34. 如权利要求28所述的CT内部感兴趣区域成像方法，其特征在于：所述交替进行感兴趣区域图像重建和感兴趣区域图像修复包括：
    循环执行所述依据内部感兴趣区域图像重建模型进行感兴趣区域图像重建和所述依据内部感兴趣区域图像修复模型进行感兴趣区域图像修复；
    当循环次数达到预设的次数时，停止迭代运算，并将所得到的迭代运算结果作为所获得的感兴趣区域重建图像。
  35. 一种CT内部感兴趣区域成像系统，其特征在于，包括：
    扫描子系统，用于扫描内部感兴趣区域，获取通过内部感兴趣区域的所有射线投影以及重建系统参数；
    图像重建子系统，用于依据内部感兴趣区域图像重建模型，进行内部感兴趣区域图像重建；
    图像修复子系统，用于依据内部感兴趣区域图像修复模型，进行内部感兴趣区域图像修复；
    迭代子系统，用于判断所述图像重建子系统和图像修复子系统交替进行感兴趣区域图像重建和感兴趣区域图像修复是否已经达到预定的交替迭代次数；以及用于输出感兴趣区域重建图像。
PCT/CN2015/075669 2014-12-30 2015-04-01 Ct成像方法和系统 Ceased WO2016106990A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201410849715.4A CN104637033B (zh) 2014-12-30 2014-12-30 Ct内部感兴趣区域成像方法和系统
CN201410849715.4 2014-12-30

Publications (1)

Publication Number Publication Date
WO2016106990A1 true WO2016106990A1 (zh) 2016-07-07

Family

ID=53215740

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2015/075669 Ceased WO2016106990A1 (zh) 2014-12-30 2015-04-01 Ct成像方法和系统

Country Status (2)

Country Link
CN (1) CN104637033B (zh)
WO (1) WO2016106990A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109389575A (zh) * 2018-10-09 2019-02-26 山东理工大学 一种基于代数迭代法的图像快速局部重建方法
CN110458908A (zh) * 2019-08-05 2019-11-15 赛诺威盛科技（北京）有限公司 基于有限角度迭代重建超视野ct图像的方法
CN110599563A (zh) * 2019-09-11 2019-12-20 河南工业大学 一种自适应nlm修正的ct重建方法及装置
CN111968192A (zh) * 2020-06-29 2020-11-20 深圳先进技术研究院 一种ct图像的构建方法、ct设备以及存储介质
CN112652031A (zh) * 2020-12-30 2021-04-13 上海联影智能医疗科技有限公司 伪影校正分析方法、电子设备及存储介质

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106023116B (zh) * 2016-05-30 2018-09-28 中国科学院深圳先进技术研究院 一种基于块加权约束的压缩感知图像重建方法及装置
CN106691485A (zh) * 2016-07-18 2017-05-24 山东省肿瘤防治研究院 一种放射影像引导下诊治肺肿瘤装置
CN107886478B (zh) * 2017-09-22 2020-10-30 深圳先进技术研究院 一种ct图像重建方法及系统、终端及可读存储介质
CN107845120B (zh) * 2017-09-27 2020-11-27 深圳先进技术研究院 Pet图像重建方法、系统、终端和可读存储介质
CN114758079B (zh) * 2022-05-19 2024-02-20 国网河南省电力公司电力科学研究院 电气设备内部结构局部成像三维重建方法
CN118447113B (zh) * 2024-04-24 2025-02-25 清华大学 扩大成像视野的x射线ct局部成像方法、装置及电子设备

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103514615A (zh) * 2012-06-22 2014-01-15 通用电气公司 用于迭代重建的方法和设备
CN104217448A (zh) * 2014-09-05 2014-12-17 中国科学院深圳先进技术研究院 基于迭代特征修正的磁共振快速成像方法及系统

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8135186B2 (en) * 2008-01-25 2012-03-13 Purdue Research Foundation Method and system for image reconstruction
JP5662447B2 (ja) * 2009-08-20 2015-01-28 コーニンクレッカ フィリップス エヌ ヴェ 関心領域画像の再構成
CN103606177B (zh) * 2013-11-26 2016-08-17 南方医科大学 稀疏角度的ct图像迭代重建方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103514615A (zh) * 2012-06-22 2014-01-15 通用电气公司 用于迭代重建的方法和设备
CN104217448A (zh) * 2014-09-05 2014-12-17 中国科学院深圳先进技术研究院 基于迭代特征修正的磁共振快速成像方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
LIU, WENLEI ET AL.: "Image Reconstruction for Low-Dose CT and Sparse-View", CHINA MASTER'S THESES FULL-TEXT DATABASE (ELECTRONIC JOURNALS, 31 March 2014 (2014-03-31), pages 1 and 33 - 38 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109389575A (zh) * 2018-10-09 2019-02-26 山东理工大学 一种基于代数迭代法的图像快速局部重建方法
CN109389575B (zh) * 2018-10-09 2023-12-01 山东理工大学 一种基于代数迭代法的图像快速局部重建方法
CN110458908A (zh) * 2019-08-05 2019-11-15 赛诺威盛科技（北京）有限公司 基于有限角度迭代重建超视野ct图像的方法
CN110458908B (zh) * 2019-08-05 2022-12-09 赛诺威盛科技(北京)股份有限公司 基于有限角度迭代重建超视野ct图像的方法
CN110599563A (zh) * 2019-09-11 2019-12-20 河南工业大学 一种自适应nlm修正的ct重建方法及装置
CN110599563B (zh) * 2019-09-11 2022-10-18 河南工业大学 一种自适应nlm修正的ct重建方法及装置
CN111968192A (zh) * 2020-06-29 2020-11-20 深圳先进技术研究院 一种ct图像的构建方法、ct设备以及存储介质
CN111968192B (zh) * 2020-06-29 2024-10-22 深圳先进技术研究院 一种ct图像的构建方法、ct设备以及存储介质
CN112652031A (zh) * 2020-12-30 2021-04-13 上海联影智能医疗科技有限公司 伪影校正分析方法、电子设备及存储介质

Also Published As

Publication number Publication date
CN104637033B (zh) 2018-02-06
CN104637033A (zh) 2015-05-20

Similar Documents

Publication Publication Date Title
WO2016106990A1 (zh) Ct成像方法和系统
CN109685206B (zh) 生成用于图像处理的神经网络模型的系统和方法
EP3685350B1 (en) Image reconstruction using machine learning regularizers
Xu et al. A practical cone-beam CT scatter correction method with optimized Monte Carlo simulations for image-guided radiation therapy
Yu et al. A soft-threshold filtering approach for reconstruction from a limited number of projections
JP7041128B6 (ja) ノイズに起因するアーチファクト形成の動的抑制を用いた反復画像再構成
Lauzier et al. Characterization of statistical prior image constrained compressed sensing (PICCS): II. Application to dose reduction
CN113516586B (zh) 一种低剂量ct图像超分辨率去噪方法和装置
WO2020151424A1 (zh) 一种基于各向异性全变分的有限角度ct重建算法
CN110660123A (zh) 基于神经网络的三维ct图像重建方法和设备以及存储介质
US11276208B2 (en) System, method and computer-accessible medium for ultralow dose computed tomography image reconstruction
Wang et al. Low-dose spectral CT reconstruction based on image-gradient L0-norm and adaptive spectral PICCS
CN104240270A (zh) Ct成像方法和系统
Bao et al. Few‐view CT reconstruction with group‐sparsity regularization
CN103106676B (zh) 一种基于低剂量投影数据滤波的x射线ct图像重建方法
JP2019524356A (ja) 異なる反復から抽出された特徴画像を使用する特徴ベースの画像処理
CN103186882A (zh) Pet系统中图像的衰减校正方法及装置
Friot--Giroux et al. Iterative tomographic reconstruction with TV prior for low-dose CBCT dental imaging
Sharma et al. Interior micro‐CT with an offset detector
Chen et al. CT metal artifact reduction method based on improved image segmentation and sinogram in‐painting
JP2021065707A (ja) 医用画像処理装置、学習済みモデルおよび医用画像処理方法
Li et al. Sparse CT reconstruction based on multi-direction anisotropic total variation (MDATV)
Maier et al. GPU denoising for computed tomography
Zhu et al. Iterative CT reconstruction via minimizing adaptively reweighted total variation
Arya et al. Regularization based modified SART iterative method for CT image reconstruction

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: 15874683

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 30.10.17)

122 Ep: pct application non-entry in european phase

Ref document number: 15874683

Country of ref document: EP

Kind code of ref document: A1