• Corpus ID: 53652346

gpuNUFFT - An Open Source GPU Library for 3D Regridding with Direct Matlab Interface

  title={gpuNUFFT - An Open Source GPU Library for 3D Regridding with Direct Matlab Interface},
  author={Florian Knoll and Andreas Schwarzl and Clemens Diwoky and Daniel K. Sodickson},
Target Audience: Researchers and clinicians interested in 3D non-Cartesian image reconstruction. Purpose: 3D non-Cartesian trajectories are very attractive for iterative image reconstruction [1], and especially in compressed sensing [2], because of incoherent aliasing in the case of undersampling. However, image reconstruction is still challenging due to prohibitively expensive computation times. The computational bottlenecks are usually gridding and inverse gridding, which have to be evaluated… 

Figures from this paper

gNUFFTW : Auto-Tuning for High-Performance GPU-Accelerated Non-Uniform Fast Fourier Transforms by Teresa
  • T. Ou
  • Computer Science
  • 2017
An empirical auto-tuning approach for the non-uniform Fast Fourier Transform on General Purpose Graphics Processors Units (GPGPU) is demonstrated and order-of-magnitude speed improvements with autotuning are demonstrated compared to typical default choices.
cuFINUFFT: a load-balanced GPU library for general-purpose nonuniform FFTs
This work presents a general-purpose GPU-based CUDA library for type 1 (nonuniform to uniform) and type 2 ( uniform to nonuniform) transforms in dimensions 2 and 3, in single or double precision, and demonstrates a 5–12× speedup versus CPU in an X-ray diffraction 3D iterative reconstruction task at 10−12 accuracy.
3807 NUFFT : Fast Auto-Tuned GPU-Based Library
A fast auto-tuned l ibrary for computing non-uniform fast Fourier Transform (NUFFT) on GPU, which includes forward and adjoint NUFFT using precomputation-free and ful ly-precomputed methods, and can be extended to include implementations from other l ibraries.
Rapid dealiasing of undersampled, non‐Cartesian cardiac perfusion images using U‐net
This study shows that the proposed pipeline with a U‐Net is capable of reconstructing 6.4‐fold accelerated, non‐Cartesian cardiac perfusion k‐space data at least 10 times faster than CS, without significant loss in data fidelity or image quality.
Python Non-Uniform Fast Fourier Transform (PyNUFFT): An Accelerated Non-Cartesian MRI Package on a Heterogeneous Platform (CPU/GPU)
A Python non-uniform fast Fourier transform (PyNUFFT) package has been developed to accelerate multidimensional non-Cartesian image reconstruction on heterogeneous platforms and provides several solvers, including the conjugate gradient method, l1 total variation regularized ordinary least square (L1TV-OLS), and l1total variation regularization least absolute deviation (L 1TV-LAD).
Memory reduced non-Cartesian MRI encoding using the mixed-radix tensor product on CPU and GPU
This work reformulates the multi-dimensional non-uniform fast Fourier transform (NUFFT) to a tensor form and proposes a tree-like, mixed-radix tensor method which flexibly reduces the storage of the NUFFT.
Optimizing full 3D SPARKLING trajectories for high-resolution T2*-weighted Magnetic Resonance Imaging
It is shown that this novel optimization for full 3D SPARKLING outperforms stacking strategies or 3D twisted projection imaging through retrospective and prospective studies on NIST phantom and in vivo brain scans at 3 Tesla without compromising image quality.
A parallel non-uniform fast Fourier transform library based on an "exponential of semicircle" kernel
FINUFFT is presented, an efficient parallel library for type 1 (nonuiform to uniform), type 2 (uniform to nonuniform), or type 3 (non uniform toNonuniform) transforms, in dimensions 1, 2, or 3, which uses minimal RAM, requires no precomputation or plan steps, and has a simple interface to several languages.
Clinically Deployed Distributed Magnetic Resonance Imaging Reconstruction: Application to Pediatric Knee Imaging
This work analyzes the practical limitations to T2 Shuffling, a four-dimensional CS-based acquisition, which provides sharp 3D-isotropic-resolution and multi-contrast images in a single scan, and reduces the reconstruction time in the hospital to 90 seconds on a 4-node cluster, enabling its use clinically.
Accelerating compressed sensing reconstruction of subsampled radial k-space data using geometrically-derived density compensation
The proposed g DCF accelerates CS reconstruction of subsampled radial k-space data without significant loss in image quality compared with no DCF as the reference, confirming that the gDCF produces uniform densities at the upper limit (Nyquist).