WO2021184469A1 - 一种放射状采集扩散加权成像运动伪影校正方法 - Google Patents

一种放射状采集扩散加权成像运动伪影校正方法 Download PDF

Info

Publication number
WO2021184469A1
WO2021184469A1 PCT/CN2020/084732 CN2020084732W WO2021184469A1 WO 2021184469 A1 WO2021184469 A1 WO 2021184469A1 CN 2020084732 W CN2020084732 W CN 2020084732W WO 2021184469 A1 WO2021184469 A1 WO 2021184469A1
Authority
WO
WIPO (PCT)
Prior art keywords
projection data
damaged
original
projection
data
Prior art date
Application number
PCT/CN2020/084732
Other languages
English (en)
French (fr)
Inventor
吴子岳
罗海
陈潇
Original Assignee
无锡鸣石峻致医疗科技有限公司
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 无锡鸣石峻致医疗科技有限公司 filed Critical 无锡鸣石峻致医疗科技有限公司
Publication of WO2021184469A1 publication Critical patent/WO2021184469A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/80—Geometric correction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10072—Tomographic images
    • G06T2207/10088—Magnetic resonance imaging [MRI]

Definitions

  • the invention relates to the field of nuclear magnetic resonance imaging, in particular to a method for correcting motion artifacts in radial acquisition diffusion weighted imaging.
  • Diffusion Weighted Imaging is an imaging method that non-invasively reflects the diffusion motion of water molecules in a living body at the molecular level, and is currently the only imaging method for measuring the diffusion motion of water molecules in a living body.
  • Diffusion-weighted imaging mainly relies on the movement of water molecules rather than the proton density of the tissue, T1 or T2 relaxation time.
  • Diffusion-weighted imaging is suitable for detecting the micro-dynamics and micro-structural changes of biological tissues at the level of living cells, and plays a pivotal role in the identification of benign and malignant tumors, evaluation of curative effects, and prediction.
  • single-shot Echo Planar Imaging EPI
  • the characteristic of single-shot EPI imaging is that the scanning time is short and the subject's motion is less affected.
  • Single shot imaging technology also has its own shortcomings. First, due to the small acquisition bandwidth along the phase encoding direction, serious image distortion will occur at the junction of different tissues with large differences in magnetic media rates; second, when acquiring high-resolution images with a single excitation, you need to Long echo chain, long echo chain means a large T2 attenuation, which will cause image blur and greatly reduce the sexual-to-noise ratio.
  • multiple excitation diffusion-weighted imaging In order to reduce image distortion and improve image resolution and signal-to-noise ratio, multiple excitation diffusion-weighted imaging has become a new research hotspot in recent years. Based on the different acquisition methods, multiple excitation diffusion weighted imaging is mainly divided into multiple excitation EPI diffusion weighted imaging, multiple excitation spiral acquisition trajectory (Spiral) diffusion weighted imaging, and multiple excitation Fast Spin Echo (FSE) ) Diffusion-weighted imaging and multiple-shot radial acquisition diffusion-weighted imaging. The first three imaging methods all require complex algorithms to deal with the phase error generated between multiple excitations, the reconstruction time is slow, and the stability is poor. Diffusion-weighted imaging based on radial acquisition can convert K-space data into projection data through one-dimensional Fourier transform.
  • the projection data is modeled, it is reconstructed by computer tomography (CT) reconstruction technology, such as filtering.
  • CT computer tomography
  • FBP Filtered Back Projection
  • FIG. 1 is the diffusion-weighted projection data of volunteers' heads collected radially. Each row of data is the projection data collected by one excitation. It can be seen that there are multiple rows of data with varying degrees of damage due to movement out of phase.
  • the invention aims to provide a method for correcting motion artifacts in radial acquisition diffusion weighted imaging, which can reduce motion artifacts and improve image quality.
  • the invention discloses a method for correcting motion artifacts in radial acquisition diffusion weighted imaging, which includes the following steps:
  • S700 Perform filtering back projection transformation on the restored projection data set P c to obtain a final reconstructed image.
  • step S500 the calculation formula of S is
  • T 0.2.
  • step S600 the calculation formula of P c is
  • ⁇ is the weight corresponding to the distance between the projection data point and the nearest damaged data point along the one-dimensional direction of a single piece of projection data.
  • ⁇ d/N, where d is the distance between the projected data point and the nearest damaged data point.
  • N 5.
  • the present invention proposes a correction method by detecting and repairing damaged data, which can reduce the motion artifacts and improve the image quality.
  • Figure 1 shows the radial diffusion weighted projection data before restoration
  • Figure 2 is a schematic flow diagram of the present invention
  • Figure 3 shows the radial diffusion weighted projection data after restoration.
  • K-space K-space, the frequency domain space of magnetic resonance signals
  • DWI Diffusion Weighted Imaging, Diffusion Weighted Imaging or Diffusion Weighted Imaging
  • Multi-Shot multiple shots
  • Multi-Shot DWI Multi-shot diffusion weighted imaging
  • T1 Time constant for regrowth of longitudinal magnetization after RF-pulse, longitudinal magnetization vector recovery time constant
  • T2 Time constant for decay of transverse magnetization after RF-pulse, transverse magnetization vector decay time constant
  • TR Repetition Time, repetition time or repetition period
  • EPI Echo planar imaging, planar echo imaging technology
  • Radial imaging Radial K-space imaging technology (traditional magnetic resonance imaging is based on Cartesian K-space acquisition technology)
  • CT Computer Tomography, computer tomography
  • FBP Filtered Back Projection, filtered back projection reconstruction
  • the present invention includes the following steps:
  • Step 1 Collect radial K-space data
  • Step 2 Perform one-dimensional Fourier transform along the readout direction and take the modulus to obtain the original radial projection data set P raw , which contains random damaged data;
  • Step 3 Based on the original projection data set P raw , the original image M raw is obtained using filtered back projection reconstruction;
  • Step 4 Based on the original image M raw , a new projection data set P new is obtained through Radon transformation;
  • Step 5 Compare the original projection data set P raw and the new projection data set P new to obtain the damaged data mask area S,
  • T is a threshold, which can be set through experience.
  • the sixth step is data fusion.
  • the projection data of the damaged area is covered with a new projection data set, and the points adjacent to the damaged area are linearly fused with the original projection data and the new projection data to achieve a smooth transition.
  • ⁇ is the weight corresponding to the distance between the projection data point and the nearest damaged data point along the one-dimensional direction of a single piece of projection data.
  • ⁇ d/N
  • d is the distance between the projection data point and the nearest damaged data point
  • the distance is in pixels.
  • the value of N in the formula can be 5, but it is not limited to this;
  • the seventh step is to perform filtering back projection transformation on the restored projection data set P c to obtain the final reconstructed image.

Landscapes

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

Abstract

本发明公开一种放射状采集扩散加权成像运动伪影校正方法，包括受损数据检测、受损数据修复和滤波反投影重建。受损数据检测步骤包括，将原始K空间数据转换到投影数据空间得到原始投影数据集，将原始投影数据集由滤波反投影重建出原始图像，接着将原始图像通过Radon变换得到新的投影数据集，通过比较新的投影数据集和原始投影数据集的差异，检测出运动受损数据。受损数据修复包括，将受损区域的数据用新的投影数据替换，将受损区域边缘的数据用原始投影数据和新的投影数据进行线性融合，得到校正的投影数据。最后，利用滤波反投影重建将校正的投影数据重建出最终图像。本发明方案可以减轻运动伪影，提升图像质量。

Description

一种放射状采集扩散加权成像运动伪影校正方法 技术领域
本发明涉及核磁共振成像领域，尤其涉及一种放射状采集扩散加权成像运动伪影校正方法。
背景技术
扩散加权成像(Diffusion Weighted Imaging,DWI)是一种在分子水平上无创地反映活体水分子扩散运动的成像方法，是目前活体测量水分子扩散运动的唯一影像手段。扩散加权成像主要依赖于水分子的运动而非组织的质子密度、T1或T2弛豫时间。扩散加权成像适用于活体细胞水平探测生物组织的微动态和微结构变化，在肿瘤的良恶性鉴别、疗效评估和预测起着举足轻重的作用。
目前临床上广泛使用的扩散成像方法通常是单次激发平面回波成像(Echo Planar Imaging，简称EPI)。单次激发EPI成像的特点是扫描时间短，受被试者的运动影响较小。单次激发成像技术也有它本身的不足。第一，由于沿着相位编码方向的采集带宽较小，在磁介质率相差较大的不同组织交界处会产生较严重的图像变形；第二，单次激发方式获取高分辨率图像时，需要长回波链，长回波链意味着大的T2衰减，会导致图像模糊以及极大降低性噪比。
为了减小图像变形，提高图像分辨率和信噪比，近年来多次激发扩散加权成像成为新的研究热点。基于采集方式的不同，多次激发扩散加权成像主要分为多次激发EPI扩散加权成像，多次激发螺旋状采集轨迹(Spiral)扩散加权成像， 多次激发快速自旋回波(Fast Spin Echo,FSE)扩散加权成像以及多次激发放射状采集扩散加权成像。前三种成像方式，都需要复杂的算法来处理多次激发之间产生的相位误差，重建时间慢，稳定性差。基于放射状采集的扩散加权成像，可以通过一维傅里叶变换，将K空间数据转换为投影数据，对投影数据取模后，利用计算机断层成像(computer tomography，CT)重建技术进行重建，如滤波反投影重建(Filtered Back Projection，FBP)算法重建图像，从而完全不用考虑多次激发之间相位误差的影响，算法更快更稳定，具有很大的应用潜力。
在扩散加权成像中，施加了巨大的运动敏感梯度，一方面可以因水分子扩散速率不同形成扩散加权对比度，另一方面会对各种宏观运动非常敏感而导致运动伪影。例如病人自主运动，生理运动以及机械振动等，会导致质子无法完全聚相，从而随机的破坏原始数据，形成伪影。图1即为基于放射状采集的志愿者头部扩散加权投影数据，每一行数据为一次激发采集到的投影数据，可见有多行数据存在不同程度的破损，为运动失相所致。
发明内容
本发明旨在提供一种放射状采集扩散加权成像运动伪影校正方法，可以减轻运动伪影，提升图像质量。
为达到上述目的，本发明是采用以下技术方案实现的：
本发明公开一种放射状采集扩散加权成像运动伪影校正方法，包括以下步骤：
S100、采集放射状K空间数据；
S200、沿读出方向进行一维傅里叶变换，并取模值，得到原始放射状投影数据集P raw，该数据集存在随机的受损数据；
S300、基于原始投影数据集P raw，利用滤波反投影重建获得原始图像M raw；
S400、基于原始图像M raw，通过Radon变换，得到新的投影数据集P new；
S500、比较原始投影数据集P raw和新的投影数据集P new，得到受损数据掩码区域S；
S600、数据融合：将受损区域的投影数据，用新的投影数据集覆盖，而与受损区域相邻的点，用原始投影数据和新的投影数据线性融合，达到平滑过渡，最终得到修复后的投影数据集P c；
S700、对修复后的投影数据集P c进行滤波反投影变换，得到最终的重建图像。
优选的，步骤S500中，S的计算公式为
Figure PCTCN2020084732-appb-000001
优选的，T＝0.2。
优选的，步骤S600中，P c的计算公式为
Figure PCTCN2020084732-appb-000002
其中α为沿单条投影数据一维方向，该投影数据点离最近的受损数据点的距离对应的权重。
优选的，α＝d/N,d为该投影数据点离最近的受损数据点的距离。
优选的，N＝5。
本发明的有益效果：
本发明针对放射状采集的扩散加权成像中存在的运动伪影，提出了一种通过检测并修复受损数据的校正方法，可以减轻运动伪影，提升图像质量。
附图说明
图1为修复前的放射状扩散加权投影数据；
图2为本发明的流程示意图；
图3为修复后的放射状扩散加权投影数据。
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，对本发明进行进一步详细说明。
本申请中：
K-space：K空间，磁共振信号的频域空间
DWI：Diffusion Weighted Imaging，弥散加权成像或扩散加权成像
Multi-Shot：多次激发
Multi-Shot DWI：多次激发扩散加权成像
T1：Time constant for regrowth of longitudinal magnetization after RF-pulse，纵向磁化矢量恢复时间常数
T2：Time constant for decay of transverse magnetization after RF-pulse，横向磁化矢量衰减时间常数
TR：Repetition Time，重复时间或重复周期
EPI：Echo planar imaging，平面回波成像技术
Radial imaging：放射状K空间成像技术(传统磁共振成像基于笛卡尔K空间采集技术)
CT：Computer Tomography，计算机断层成像
FBP：Filtered Back Projection，滤波反投影重建
如图2、图3所示，本发明包括以下步骤：
第1步：采集放射状K空间数据；
第2步：沿读出方向进行一维傅里叶变换，并取模值，得到原始放射状投影数据集P raw，该数据集存在随机的受损数据；
第3步：基于原始投影数据集P raw，利用滤波反投影重建获得原始图像M raw；
第4步：基于原始图像M raw，通过Radon变换，得到新的投影数据集P new；
第5步：比较原始投影数据集P raw和新的投影数据集P new，得到受损数据掩码区域S，
具体的，S通过如下公式获得
Figure PCTCN2020084732-appb-000003
其中T为一个阈值，可以通过经验设定。优选的，T＝0.2可以取得较好结果，但不限于此；
第6步，数据融合，将受损区域的投影数据，用新的投影数据集覆盖，而与受损区域相邻的点，用原始投影数据和新的投影数据线性融合，达到平滑过渡，最终得到修复后的投影数据集P c，具体的
Figure PCTCN2020084732-appb-000004
其中，α为沿单条投影数据一维方向，该投影数据点离最近的受损数据点的距离对应的权重。本实施例中，α＝d/N,d为该投影数据点离最近的受损数据点的距离，该距离以像素为单位，该式中N的值可以取5，但不限于此；
第7步，对修复后的投影数据集P c进行滤波反投影变换，得到最终的重建图像。
当然，本发明还可有其它多种实施例，在不背离本发明精神及其实质的情况下，熟悉本领域的技术人员可根据本发明作出各种相应的改变和变形，但这些相应的改变和变形都应属于本发明所附的权利要求的保护范围。

Claims (6)

  1. 一种放射状采集扩散加权成像运动伪影校正方法，其特征在于包括以下步骤：
    S100、采集放射状K空间数据；
    S200、沿读出方向进行一维傅里叶变换，并取模值，得到原始放射状投影数据集P raw，该数据集存在随机的受损数据；
    S300、基于原始投影数据集P raw，利用滤波反投影重建获得原始图像M raw；
    S400、基于原始图像M raw，通过Radon变换，得到新的投影数据集P new；
    S500、比较原始投影数据集P raw和新的投影数据集P new，得到受损数据掩码区域S；
    S600、数据融合：将受损区域的投影数据，用新的投影数据集覆盖，而与受损区域相邻的点，用原始投影数据和新的投影数据线性融合，达到平滑过渡，最终得到修复后的投影数据集P c；
    S700、对修复后的投影数据集P c进行滤波反投影变换，得到最终的重建图像。
  2. 根据权利要求1所述的校正方法，其特征在于：步骤S500中，S的计算公式为
    Figure PCTCN2020084732-appb-100001
  3. 根据权利要求2所述的校正方法，其特征在于：T＝0.2。
  4. 根据权利要求2或3所述的校正方法，其特征在于：步骤S600中，P c的计算公式为
    Figure PCTCN2020084732-appb-100002
    其中α为沿单条投影数据一维方向，该投影数据点离最近的受损数据点的距离对应的权重。
  5. 根据权利要求4所述的校正方法，其特征在于：α＝d/N,d为该投影数据点离最近的受损数据点的距离。
  6. 根据权利要求5所述的校正方法，其特征在于：N＝5。
PCT/CN2020/084732 2020-03-17 2020-04-14 一种放射状采集扩散加权成像运动伪影校正方法 WO2021184469A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010188632.0A CN111415315B (zh) 2020-03-17 2020-03-17 一种放射状采集扩散加权成像运动伪影校正方法
CN202010188632.0 2020-03-17

Publications (1)

Publication Number Publication Date
WO2021184469A1 true WO2021184469A1 (zh) 2021-09-23

Family

ID=71494302

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/084732 WO2021184469A1 (zh) 2020-03-17 2020-04-14 一种放射状采集扩散加权成像运动伪影校正方法

Country Status (2)

Country Link
CN (1) CN111415315B (zh)
WO (1) WO2021184469A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113256684B (zh) * 2021-06-11 2021-10-01 腾讯科技（深圳）有限公司 动捕数据的修复方法、装置、设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060239585A1 (en) * 2005-04-04 2006-10-26 Valadez Gerardo H System and method for reducing artifacts in motion corrected dynamic image sequences
CN104952039A (zh) * 2015-06-14 2015-09-30 中国电子科技集团公司第十研究所 分布式图像压缩感知重建方法
CN110652296A (zh) * 2019-09-16 2020-01-07 华东师范大学 一种去除磁共振头部图像运动伪影的方法

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102005031B (zh) * 2010-11-03 2012-05-02 宁波鑫高益磁材有限公司 一种mri系统中k空间采样数据的运动伪影消除方法及装置
US9615812B2 (en) * 2013-06-19 2017-04-11 Koninklijke Philips N.V. Calibration of imagers with dynamic beam shapers
CN105469366B (zh) * 2015-11-23 2018-09-04 山东科技大学 一种消减ct图像金属伪影的解析方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060239585A1 (en) * 2005-04-04 2006-10-26 Valadez Gerardo H System and method for reducing artifacts in motion corrected dynamic image sequences
CN104952039A (zh) * 2015-06-14 2015-09-30 中国电子科技集团公司第十研究所 分布式图像压缩感知重建方法
CN110652296A (zh) * 2019-09-16 2020-01-07 华东师范大学 一种去除磁共振头部图像运动伪影的方法

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
HAN YEJI, HWANG JINYOUNG, CHUNG JUN-YOUNG, YUN SUNGDAE, PARK HYUNWOOK: "Improvement of the Diffusion-Weighted Images Acquired With Radial Trajectories Using Projection Data Regeneration", JOURNAL OF MAGNETIC RESONANCE IMAGING, vol. 26, no. 3, 1 September 2007 (2007-09-01), US, pages 799 - 804, XP055851726, ISSN: 1053-1807, DOI: 10.1002/jmri.21057 *
ZHAO JIN-LI , HU HONG-PING , LI QUAN: "Projection Fused Image Reconstruction Algorithm Study Based on Graded Variable Voltage", NUCLEAR ELECTRONICS & DETECTION TECHNOLOGY, vol. 35, no. 6, 20 June 2015 (2015-06-20), pages 585 - 589, XP055851730, ISSN: 0258-0934 *
ZHAO JINLI: "The Study of CT Image Fusion And Reconstruction Based on Graded Variable Voltage", CHINESE MASTER’S THESES FULL-TEXT DATABASE, 1 June 2016 (2016-06-01), pages 1 - 56, XP055851735, ISSN: 1674-0246 *

Also Published As

Publication number Publication date
CN111415315A (zh) 2020-07-14
CN111415315B (zh) 2023-09-26

Similar Documents

Publication Publication Date Title
Hauptmann et al. Real‐time cardiovascular MR with spatio‐temporal artifact suppression using deep learning–proof of concept in congenital heart disease
Lyu et al. Cine cardiac MRI motion artifact reduction using a recurrent neural network
Cordero-Grande et al. Sensitivity encoding for aligned multishot magnetic resonance reconstruction
US8427153B2 (en) Method for motion correction in magnetic resonance imaging using radio frequency coil arrays
US7358730B2 (en) Diffusion tensor imaging using highly constrained image reconstruction method
US9395431B2 (en) Multi-contrast delayed enhancement cardiac magnetic resonance imaging
US20230360179A1 (en) Machine learning processing of contiguous slice image data
Yang et al. First‐pass myocardial perfusion imaging with whole‐heart coverage using L1‐SPIRiT accelerated variable density spiral trajectories
Qi et al. End‐to‐end deep learning nonrigid motion‐corrected reconstruction for highly accelerated free‐breathing coronary MRA
Becker et al. Fast myocardial T1 mapping using cardiac motion correction
Adluru et al. Three‐dimensional late gadolinium enhancement imaging of the left atrium with a hybrid radial acquisition and compressed sensing
Naresh et al. Accelerated, first‐pass cardiac perfusion pulse sequence with radial k‐space sampling, compressed sensing, and k‐space weighted image contrast reconstruction tailored for visual analysis and quantification of myocardial blood flow
CN109696647A (zh) 三维多次激发扩散加权成像的k空间采集方法及重建方法
Johansson et al. Abdominal DCE‐MRI reconstruction with deformable motion correction for liver perfusion quantification
WO2021184469A1 (zh) 一种放射状采集扩散加权成像运动伪影校正方法
Li et al. High‐resolution diffusion‐weighted imaging with interleaved variable‐density spiral acquisitions
Ljungberg et al. Motion corrected silent ZTE neuroimaging
JP4478694B2 (ja) 非線形位相補正方法、記録媒体および磁気共鳴装置
Usman et al. Motion corrected multishot MRI reconstruction using generative networks with sensitivity encoding
Lugauer et al. Single-breath-hold abdominal T _ 1 T 1 mapping using 3D Cartesian Look-Locker with spatiotemporal sparsity constraints
US20230380714A1 (en) Method and system for low-field mri denoising with a deep complex-valued convolutional neural network
Delzescaux et al. Using an adaptive semiautomated self‐evaluated registration technique to analyze MRI data for myocardial perfusion assessment
WO2022217157A1 (en) System and method for quantitative magnetic resonance imaging using a deep learning network
Awate et al. Model-based image reconstruction for dynamic cardiac perfusion MRI from sparse data
Eichhorn et al. Deep Learning-Based Detection of Motion-Affected k-Space Lines for T2*-Weighted MRI

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: 20925278

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: 20925278

Country of ref document: EP

Kind code of ref document: A1