检索结果-内蒙古大学图书馆

FlowMRI-Net: A Generalizable Self-Supervised 4D Flow MRI Reconstruction Network

Journal of Cardiovascular Magnetic Resonance 2025年 101913页

作者： Luuk Jacobs Marco Piccirelli Valery Vishnevskiy Sebastian Kozerke Institute for Biomedical Engineering University and ETH Zurich Zurich Switzerland Department of Neuroradiology University Hospital Zurich Zurich Switzerland

Background Image reconstruction from highly undersampled 4D flow MRI data can be very time consuming and may result in significant underestimation of velocities depending on regularization, thereby limiting the applicability of the method. The objective of the present work was to develop a generalizable self-supervised deep learning-based framework for fast and accurate reconstruction of highly undersampled 4D flow MRI and to demonstrate the utility of the framework for aortic and cerebrovascular applications. Methods The proposed deep-learning-based framework, called FlowMRI-Net, employs physics-driven unrolled optimization using a complex-valued convolutional recurrent neural network and is trained in a self-supervised manner. The generalizability of the framework is evaluated using aortic and cerebrovascular 4D flow MRI acquisitions acquired on systems from two different vendors for various undersampling factors (R=8,16,24) and compared to compressed sensing (CS-LLR) reconstructions. Evaluation includes an ablation study and a qualitative and quantitative analysis of image and velocity magnitudes. Results FlowMRI-Net outperforms CS-LLR for aortic 4D flow MRI reconstruction, resulting in significantly lower vectorial normalized root mean square error and mean directional errors for velocities in the thoracic aorta. Furthermore, the feasibility of FlowMRI-Net’s generalizability is demonstrated for cerebrovascular 4D flow MRI reconstruction. Reconstruction times ranged from 3 to 7 minutes on commodity CPU/GPU hardware. Conclusion FlowMRI-Net enables fast and accurate reconstruction of highly undersampled aortic and cerebrovascular 4D flow MRI, with possible applications to other vascular territories.

关键词： 2D two-dimensional 3D three-dimensional 4D four-dimensional 5D five-dimensional MRI magnetic resonance imaging PI parallel imaging CS compressed sensing LLR locally low rank REF reference image CPU central processing unit GPU graphical processing unit DL deep learning CNN convolutional neural network RNN recurrent neural network VN variational network DC data consistency WA weighted averaging ECG electrocardiogram PPU peripheral pulse unit FOV field of view VENC velocity encoding TE echo time TR repetition time FA flip angle SNR signal-to-noise ratio CoW circle of Willis R undersampling factor FH feet-head RL right-left AP anteroposterior ROI region of interest MCA middle carotid artery ICA internal carotid artery PCA posterior carotid artery 4D flow MRI aorta cerebrovasculature deep learning reconstruction

来源：评论

学校读者我要写书评

暂无评论

GPU implementation of photoacoustic short-lag spatial coherence imaging for improved image-guided interventions

引用

JOURNAL OF BIOMEDICAL OPTICS 2020年第7期25卷 1-19页

作者： Gonzalez, Eduardo A. Bell, Muyinatu A. Lediju Johns Hopkins Univ Sch Med Dept Biomed Engn Baltimore MD 21205 USA Johns Hopkins Univ Dept Elect & Comp Engn Whiting Sch Engn Baltimore MD 21218 USA Johns Hopkins Univ Dept Comp Sci Whiting Sch Engn Baltimore MD 21218 USA

Significance: Photoacoustic-based visual servoing is a promising technique for surgical tool tip tracking and automated visualization of photoacoustic targets during interventional procedures. However, one outstanding challenge has been the reliability of obtaining segmentations using low-energy light sources that operate within existing laser safety limits. Aim: We developed the first known graphical processing unit (GPU)-based real-time implementation of short-lag spatial coherence (SLSC) beamforming for photoacoustic imaging and applied this real-time algorithm to improve signal segmentation during photoacoustic-based visual servoing with low-energy lasers. Approach: A 1-mm-core-diameter optical fiber was inserted into ex vivo bovine tissue. Photoacoustic-based visual servoing was implemented as the fiber was manually displaced by a translation stage, which provided ground truth measurements of the fiber displacement. GPU-SLSC results were compared with a central processing unit (CPU)-SLSC approach and an amplitude-based delay-and-sum (DAS) beamforming approach. Performance was additionally evaluated with in vivo cardiac data. Results: The GPU-SLSC implementation achieved frame rates up to 41.2 Hz, representing a factor of 348 speedup when compared with offline CPU-SLSC. In addition, GPU-SLSC successfully recovered low-energy signals (i.e., <= 268 mu J) with mean +/- standard deviation of signal-to-noise ratios of 11.2 +/- 2.4 (compared with 3.5 +/- 0.8 with conventional DAS beamforming). When energies were lower than the safety limit for skin (i.e., 394.6 mu J for 900-nm wavelength laser light), the median and interquartile range (IQR) of visual servoing tracking errors obtained with GPU-SLSC were 0.64 and 0.52 mm, respectively (which were lower than the median and IQR obtained with DAS by 1.39 and 8.45 mm, respectively). GPU-SLSC additionally reduced the percentage of failed segmentations when applied to in vivo cardiac data. Conclusions: Results are promising

关键词： coherence-based beamforming graphical processing unit photoacoustic real-time visual servoing

来源：评论

学校读者我要写书评

暂无评论

Free surface flow simulations on GPGPUs using the LBM

引用

COMPUTERS & MATHEMATICS WITH APPLICATIONS 2011年第12期61卷 3549-3563页

作者： Janssen, Christian Krafczyk, Manfred Tech Univ Carolo Wilhelmina Braunschweig Inst Computat Modeling Civil Engn Braunschweig Germany

In this paper, we present the implementation of a volume-of-fluid-(VOF)-based algorithm for the simulation of free-surface flow problems on general purpose graphical processing units (GPGPUs). For the solution of the flow field and the additional advection equation for the VOF fill level, the lattice Boltzmann method on the basis of an MRT collision operator is used. A Smagorinsky LES model serves to capture the small-scale turbulent structures of the flow. We show that despite the additional non-local operations near the phase interface, we end up with an algorithm with good overall performance, which is suitable for the simulation of demanding real-world engineering applications. (C) 2011 Elsevier Ltd. All rights reserved.

关键词： Free surface Lattice Boltzmann method D3Q19 model Volume of fluid graphical processing unit MRT Smagorinsky LES

来源：评论

学校读者我要写书评

暂无评论

Massively Parallel Approximate Gaussian Process Regression

引用

SIAM-ASA JOURNAL ON UNCERTAINTY QUANTIFICATION 2014年第1期2卷 564-584页

作者： Gramacy, Robert B. Niemi, Jarad Weiss, Robin M. Univ Chicago Booth Sch Business Chicago IL 60637 USA Iowa State Univ Dept Stat Ames IA 50011 USA Univ Chicago Ctr Res Comp Chicago IL 60637 USA

We explore how the big-three computing paradigms-symmetric multiprocessor, graphical processing units (GPUs), and cluster computing-can together be brought to bear on large-data Gaussian processes (GP) regression problems via a careful implementation of a newly developed local approximation scheme. Our methodological contribution focuses primarily on GPU computation, as this requires the most care and also provides the largest performance boost. However, in our empirical work we study the relative merits of all three paradigms to determine how best to combine them. The paper concludes with two case studies. One is a real data fluid-dynamics computer experiment which benefits from the local nature of our approximation;the second is a synthetic example designed to find the largest data set for which (accurate) GP emulation can be performed on a commensurate predictive set in under an hour.

关键词： emulator nonparametric regression graphical processing unit symmetric multiprocessor cluster computing big data computer experiment

来源：评论

学校读者我要写书评

暂无评论

Development of a real-time wave field reconstruction TEM system (I): incorporation of an auto focus tracking system

引用

MICROSCOPY 2017年第3期66卷 172-181页

作者： Tamura, Takahiro Kimura, Yoshihide Takai, Yoshizo Osaka Univ Grad Sch Engn Dept Mat & Life Sci 2-1 Yamada Oka Suita Osaka 5650871 Japan

We have developed a real-time wave field reconstruction transmission electron microscope system that enables auto focus tracking at a video rate. In the developed system, the sample drift induced in the Z-axis direction can be measured with a precision of similar to 0.4 nm and automatically corrected for every 1/30 *** have developed a real-time wave field reconstruction transmission electron microscope system that enables auto focus tracking at a video rate. In the developed system, a high-speed image calculation technique using a graphical processing unit was incorporated along with two techniques facilitating high-speed focus control using high-voltage modulation and weighted image integration using exposure time control. By utilizing these techniques, the sample drift induced in the Z-axis direction can be measured and automatically corrected for every 1/30 of a second. The auto focus tracking system can be operated in the sample drift range of +/- 150 nm with a precision of similar to 0.4 nm. In addition to amorphous samples, the system can be used to examine thin crystalline samples, which indicates that atomic structure analysis can be performed in real time with high reliability even when heavy drift occurs in the Z-axis direction.

关键词： wave field reconstruction in situ observation real time auto focus tracking focus modulation graphical processing unit

来源：评论

学校读者我要写书评

暂无评论

GPU accelerated pivoting rules for the simplex algorithm

引用

JOURNAL OF SYSTEMS AND SOFTWARE 2014年 96卷 1-9页

作者： Ploskas, Nikolaos Samaras, Nikolaos Univ Macedonia Sch Informat Sci Dept Appl Informat Thessaloniki 54006 Greece

Simplex type algorithms perform successive pivoting operations (or iterations) in order to reach the optimal solution. The choice of the pivot element at each iteration is one of the most critical step in simplex type algorithms. The flexibility of the entering and leaving variable selection allows to develop various pivoting rules. In this paper, we have proposed some of the most well-known pivoting rules for the revised simplex algorithm on a CPU-GPU computing environment. All pivoting rules have been implemented in MATLAB and CUDA. Computational results on randomly generated optimal dense linear programs and on a set of benchmark problems (Netlib-optimal, Kennington, Netlib-infeasible, Meszaros) are also presented. These results showed that the proposed GPU implementations of the pivoting rules outperform the corresponding CPU implementations. (C) 2014 Elsevier Inc. All rights reserved.

关键词： Linear programming Simplex algorithm Pivoting rules graphical processing unit MATLAB Compute Unified Device Architecture

来源：评论

学校读者我要写书评

暂无评论

Accelerated finite element elastodynamic simulations using the GPU

引用

JOURNAL OF COMPUTATIONAL PHYSICS 2014年第parta期257卷 687-707页

作者： Huthwaite, Peter Univ London Imperial Coll Sci Technol & Med Dept Mech Engn London SW7 2AZ England

An approach is developed to perform explicit time domain finite element simulations of elastodynamic problems on the graphical processing unit, using Nvidia's CUDA. Of critical importance for this problem is the arrangement of nodes in memory, allowing data to be loaded efficiently and minimising communication between the independently executed blocks of threads. The initial stage of memory arrangement is partitioning the mesh;both a well established 'greedy' partitioner and a new, more efficient 'aligned' partitioner are investigated. A method is then developed to efficiently arrange the memory within each partition. The software is applied to three models from the fields of non-destructive testing, vibrations and geophysics, demonstrating a memory bandwidth of very close to the card's maximum, reflecting the bandwidth-limited nature of the algorithm. Comparison with Abaqus, a widely used commercial CPU equivalent, validated the accuracy of the results and demonstrated a speed improvement of around two orders of magnitude. A software package, Pogo, incorporating these developments, is released open source, downloadable from http://***/ to benefit the community. (C) 2013 Elsevier Inc. All rights reserved.

关键词： Finite element Ultrasound Elastodynamic graphical processing unit GPU

来源：评论

学校读者我要写书评

暂无评论

MRI denoising by nonlocal means on multi-GPU

引用

JOURNAL OF REAL-TIME IMAGE processing 2019年第2期16卷 523-533页

作者： Granata, Donatella Amato, Umberto Alfano, Bruno CNR Ist Applicaz Calcolo Mauro Picone Naples Italy CNR Ist Biostrutture & Bioimmagini Naples Italy

A critical issue in image restoration is noise removal, whose state-of-art algorithm, NonLocal Means, is highly demanding in terms of computational time. Aim of the present paper is to boost its performance by an efficient algorithm tailored to GPU hardware architectures. This algorithm adapts itself to several variants of the methodologies in terms of different strategies for estimating the involved filtering parameter, type of noise affecting data, multicomponent signals, spatial dimension of the images. Numerical experiments on brain Magnetic Resonance images are provided.

关键词： NonLocal means Image denoising Magnetic resonance imaging graphical processing unit

来源：评论

学校读者我要写书评

暂无评论

Highly efficient synthetic aperture radar processing system for airborne sensors using CPU plus GPU architecture

引用

JOURNAL OF APPLIED REMOTE SENSING 2015年第1期9卷 097293-1-097293-12页

作者： Wu, Zheng Liu, Yabo Zhang, Lei Li, Ning Du, Kangning Balz, Timo Chinese Acad Sci Inst Elect Space Microwave Remote Sensing Syst Dept Beijing 100190 Peoples R China Univ Chinese Acad Sci Sch Elect Elect & Commun Engn Beijing 100039 Peoples R China Wuhan Univ State Key Lab Informat Engn Surveying Mapping & R Wuhan 430079 Peoples R China

A comprehensive and effective parallel airborne synthetic aperture radar (SAR) processing system using the architecture of central processing units (CPU) and graphical processing units (GPU) is proposed first in this paper. By using the techniques of Compute Unified Device Architecture, the SAR processing system is much more efficient and robust, thereby enabling it to work with high efficiency. This paper focuses on the optimizations of motion compensation, subaperture chirp scaling algorithm, phase gradient autofocus (PGA), and even visualization with Open Source Computer Vision Library. Apart from this, it is also the first time the PGA has been run on the architecture of CPU and GPU. Experimental results show a speedup of similar to 37 times compared with a nonoptimized CPU-based approach. (C) 2015 Society of Photo-Optical Instrumentation Engineers (SPIE)

关键词： synthetic aperture radar graphical processing unit signal processing system Compute Unified Device Architecture parallel computing

来源：评论

学校读者我要写书评

暂无评论

Leveraging computation sharing and parallel processing in location-dependent query processing

引用

JOURNAL OF SUPERCOMPUTING 2012年第1期61卷 215-234页

作者： Cazalas, Jonathan Guha, Ratan Univ Cent Florida Sch Elect Engn & Comp Sci Orlando FL 32816 USA

A variety of research exists for the processing of continuous queries in large, mobile environments. Each method tries, in its own way, to address the computational bottleneck of constantly processing so many queries. In this paper, we introduce an efficient and scalable system for monitoring continuous queries by leveraging the parallel processing capability of the Graphics processing unit. We examine a naive CPU-based solution for continuous range-monitoring queries, and we then extend this system using the GPU. Additionally, with mobile communication devices becoming commodity, location-based services will become ubiquitous. To cope with the very high intensity of location-based queries, we propose a view oriented approach of the location database, thereby reducing computation costs by exploiting computation sharing amongst queries requiring the same view. Our studies show that by exploiting the parallel processing power of the GPU, we are able to significantly scale the number of mobile objects, while maintaining an acceptable level of performance.

关键词： Location-dependent query processing Mobile database systems Continuous query Location-based services graphical processing unit GPU

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：