Robust Electromagnetic Interference EMI Elimination via Simultaneous Sensing and Deep Learning Prediction for RF Shielding -free MRI Yujiao Zhao12 Linfang Xiao12 Vick Lau12

2025-04-24 1 0 9.2MB 16 页 10玖币

侵权投诉

Robust Electromagnetic Interference (EMI) Elimination via Simultaneous

Sensing and Deep Learning Prediction for RF Shielding-free MRI

Yujiao Zhao1,2, Linfang Xiao1,2, Vick Lau1,2,

Yilong Liu1,2, Alex T. Leong1,2, and Ed X. Wu1,2*

1Laboratory of Biomedical Imaging and Signal Processing

The University of Hong Kong, Hong Kong SAR, People’s Republic of China

2Department of Electrical and Electronic Engineering

The University of Hong Kong, Hong Kong SAR, People’s Republic of China

*Correspondence to:

Ed X. Wu, Ph.D.

Department of Electrical and Electronic Engineering

The University of Hong Kong, Hong Kong SAR, China

Tel: (852) 3917-7096

Fax: (852) 3917-8738

Email: ewu@eee.hku.hk

Short Running Title: Active EMI Cancellation for MRI

Keywords: MRI, RF shielding, electromagnetic interference, EMI, RF interference, RFI,

electromagnetic interference prediction, EMI prediction, electromagnetic interference removal,

EMI removal, electromagnetic interference elimination, EMI elimination, electromagnetic

interference cancellation, EMI cancellation, electromagnetic interference sensing, EMI sensing,

electromagnetic interference cancellation, EMI cancellation, deep learning

Active EMI Cancellation for MRI Zhao & Wu et al. Arxiv 2022

Page 2 of 16

ABSTRACT

At present, MRI scans are performed inside a fully-enclosed RF shielding room, posing stringent

installation requirement and unnecessary patient discomfort. We aim to develop an electromagnetic

interference (EMI) cancellation strategy for MRI with no or incomplete RF shielding. In this study,

a simultaneous sensing and deep learning driven EMI cancellation strategy is presented to model,

predict and remove EMI signals from acquired MRI signals. Specifically, during each MRI scan,

separate EMI sensing coils placed in various spatial locations are utilized to simultaneously sample

environmental and internal EMI signals within two windows (for both conventional MRI signal

acquisition and EMI characterization acquisition). Then a CNN model is trained using the EMI

characterization data to relate EMI signals detected by EMI sensing coils to EMI signals in MRI

receive coil. This model is utilized to retrospectively predict and remove EMI signals components

detected by MRI receive coil during the MRI signal acquisition window. We implemented and

demonstrated this strategy for various EMI sources on a mobile ultra-low-field 0.055 T permanent

magnet MRI scanner and a 1.5 T superconducting magnet MRI scanner with no or incomplete RF

shielding. Our experimental results demonstrate that the method is highly effective and robust in

predicting and removing various EMI sources from both external environments and internal scanner

electronics at both 0.055 T (2.3 MHz) and 1.5 T (64 MHz), producing final image signal-to-noise

ratios that are comparable to those obtained using a fully enclosed RF shielding. Our proposed

strategy enables MRI operation with no or incomplete RF shielding, alleviating MRI installation and

operational requirements. It is also potentially applicable to other scenarios of accurate RF signal

detection or discrimination in presence of external and internal EMI or RF sources.

Active EMI Cancellation for MRI Zhao & Wu et al. Arxiv 2022

Page 3 of 16

INTRODUCTION

Magnetic resonance imaging (MRI) is intrinsically superior to other medical imaging modalities (e.g.,

computed tomography and positron emission tomography), because it is non-invasive, non-ionizing,

inherently quantitative and multi-parametric1,2. However, conventional MRI scanners require

specialized and expensive installations due to infrastructural requirements, e.g., site preparation to

host the large magnets that typically weigh 3000-4500 kg, magnetic and radiofrequency (RF)

shielding, electricity to drive power-consuming electronics, and water requirement for gradient

cooling3. Consequently, the vast majority of clinical MRI scanners are housed inside fully enclosed

RF shielding room on ground floors of large hospitals and clinics, severely hindering the accessibility

and patient-friendliness of MRI in modern healthcare4.

Recently, there has been a growing impetus to develop MRI scanners at ultra-low-field (ULF)

strengths (i.e., below 0.1 Tesla or T)4-13 for low-cost clinical imaging. Preliminary results

demonstrated that these ULF MRI developments produced clinically valuable information for brain

pathology diagnosis10-15. These ULF scanners also eliminate the need for a magnetic shielding cage

because of dramatic fringe field reduction, yet many of them still require bulky and enclosed RF

shielding to prevent external electromagnetic interference (EMI) signals during data acquisition.

Such requirement precludes the portability of ULF MRI scanners for truly point-of-care applications

(e.g., in intensive care units and surgical suites).

Several solutions have been recently developed to tackle the RF shielding cage requirement for ULF

MRI9,16-20. One group used simple conductive cloth to cover the subject during scanning9. This

passive method could alter and mitigate EMI from external environments, but its performance was

suboptimal. Moreover, it was inadequate to deal with EMI from internal sources, such as scanner

console, gradient/RF amplifiers, and power supplies. Alternatively, active methods were also

proposed to reduce or eliminate external EMI. One group utilized magnetometers to sense

environmental EMI and remove EMI signals in MRI receive coil via an adaptive suppression

procedure16. This method was hardware demanding and only yielded limited success. An analytical

approach was proposed to estimate EMI signals in MRI receive coil from EMI signals detected by

EMI sensing coils using the frequency domain transfer functions among coils17. More recently, it

was extended for time domain implementation as linear convolutions and with an adaptive

procedure18. This method eliminated EMI substantially but could only produce very satisfactory brain

imaging results when used together with conductive cloth and body surface electrode for EMI pickup.

The aforementioned active EMI elimination methods are based on a simple electromagnetic

phenomenon, that is, the relationship of EMI signals detected by EMI sensing coils and MRI receive

coil can be analytically characterized by the coupling or transfer functions among coils. However, in

realistic unshielded imaging environments, EMI signals could be emitted by various sources from

external environments as well as internal scanner electronics. Further, the EMI signals can change

dynamically due to EMI sources and surrounding environments with various nature and behaviors.

Such practical issues can complicate or degrade the performance of these analytical methods.

Intuitively, a deep learning driven method is preferable over the existing analytical approaches16-20

for more robust EMI prediction and elimination. This is because neural networks potentially offer

the ability of approximating the coupling relationships among coils from a subset of nonlinear

functions, thus robustly predicting EMI signals detected by MRI receive coils especially in presence

of complex external and internal EMI signals.

Active EMI Cancellation for MRI Zhao & Wu et al. Arxiv 2022

Page 4 of 16

This study presents a novel simultaneous sensing and deep learning driven EMI cancellation strategy

for RF shielding-free MRI12,21,22. It eliminates EMI signals from acquired MRI signals by

establishing the relationships among EMI signals detected by EMI sensing coils and MRI receive

coils via deep learning. This method works effectively and robustly with regards to various and

dynamically varying EMI sources as demonstrated on our home-built mobile ULF 0.055 T head MRI

scanner and a 1.5 T whole-body clinical MRI scanner with no or incomplete RF shielding.

METHODS

EMI Cancellation via Simultaneous Sensing and Deep Learning

An EMI cancellation strategy is presented to model, predict and remove EMI signals from acquired

MRI signals by taking advantages of the well-established MRI multi-receiver electronics (previously

developed for parallel imaging) and convolutional neural networks (CNNs), as illustrated in Figure

1. EMI sensing coils are strategically placed around scanner and inside electronic cabinet to actively

detect radiative EMI signals from both external environments and internal scanner electronics

(Figure 1A). Within each TR during the scanning, main MRI receive coil and EMI sensing coils

simultaneously sample data within two acquisition windows, one is for the conventional MRI signal

acquisition, the other is chosen for acquiring the EMI characterization data (no MRI signals due to

no RF excitation, i.e., EMI signals only) (Figure 1B). After each scan, data sampled by both MRI

receive coil and EMI sensing coils within the second window (i.e., EMI characterization window)

are used to train a CNN model that can relate the 1D temporal EMI signal received by multiple EMI

sensing coils to the 1D temporal signals received by MRI receive coil for each frequency encoding

(FE) signal or k-space line (Figure 1C). This model is then applied to predict the EMI signal

component in MRI receive coil signal for each FE line within MRI signal acquisition window based

on the EMI signals simultaneously detected by EMI sensing coils. Subsequently, the predicted EMI

signal component is subtracted or removed from the MRI receive coil signals. This procedure is

repeated for all individual FE lines, creating EMI-free k-space data prior to any averaging or/and

image reconstruction.

Evaluation on a Mobile 0.055 T Head MRI Scanner

The proposed strategy was implemented and evaluated on our home-built mobile ULF 0.055 T MRI

scanner12 with no RF shielding. All experiments involving human subjects were approved by the

local institutional board and written information consents were obtained.

Phantom and brain datasets were acquired with 1 MRI receiving coil and 10 EMI sensing coils. These

EMI sensing coils were deployed around and inside the scanner to detect EMI signals that were from

both external environment and those generated internally by gradient/RF electronics during MRI

scanning. Each of them was fabricated by wounding copper wire on a 3D printed coil holder

(diameter = 5 cm) and was tuned to the Larmor frequency (2.32 MHz). The detected EMI signal was

passed through a two-stage preamplifier module (first-stage: Gain = 30 dB; second stage: Gain =30

dB, for input Vpp < 60 mV). The placement of these EMI sensing coils is depicted in Figure 1A.

Three were placed in the vicinity of the patient head holder, two on each side (left and right)

underneath the patient bed, and two in the vicinity of gradient and RF amplifiers inside the electronic

cabinet, and one underneath the scanner.

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

RobustElectromagneticInterference(EMI)EliminationviaSimultaneousSensingandDeepLearningPredictionforRFShielding-freeMRIYujiaoZhao1,2,LinfangXiao1,2,VickLau1,2,YilongLiu1,2,AlexT.Leong1,2,andEdX.Wu1,2*1LaboratoryofBiomedicalImagingandSignalProcessingTheUniversityofHongKong,HongKongSAR,People’sRepublic...

展开>> 收起<<

Robust Electromagnetic Interference EMI Elimination via Simultaneous Sensing and Deep Learning Prediction for RF Shielding -free MRI Yujiao Zhao12 Linfang Xiao12 Vick Lau12.pdf

共16页,预览4页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Robust Electromagnetic Interference EMI Elimination via Simultaneous Sensing and Deep Learning Prediction for RF Shielding -free MRI Yujiao Zhao12 Linfang Xiao12 Vick Lau12

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: