No abstract

Computational acceleration on graphics processing units (GPUs) can make advanced magnetic resonance imaging (MRI) reconstruction algorithms attractive in clinical settings, thereby improving the quality of MR images across a broad spectrum of applications. This paper describes the acceleration of such an algorithm on NVIDIA’s Quadro FX 5600. The reconstruction of a 3D image with 1283 voxels achieves up to 180 GFLOPS and requires just over one minute on the Quadro, while reconstruction on a quad-core CPU is twenty-one times slower. Furthermore, relative to the true image, the error exhibited by the advanced reconstruction is only 12%, while conventional reconstruction techniques incur error of 42%.

Supplementary data are available at Bioinformatics online.

High-resolution real-time tomography of scattering tissues is important for many areas of medicine and biology1–6. However, the compromise between transverse resolution and depth-of-field in addition to low sensitivity deep in tissue continue to impede progress towards cellular-level volumetric tomography. Computed imaging has the potential to solve these long-standing limitations. Interferometric synthetic aperture microscopy (ISAM)7–9 is a computed imaging technique enabling high-resolution volumetric tomography with spatially invariant resolution. However, its potential for clinical diagnostics remains largely untapped since full volume reconstructions required lengthy postprocessing, and the phase-stability requirements have been difficult to satisfy in vivo. Here we demonstrate how 3-D Fourier-domain resampling, in combination with high-speed optical coherence tomography (OCT), can achieve high-resolution in vivo tomography. Enhanced depth sensitivity was achieved over a depth-of-field extended in real time by more than an order of magnitude. This work lays the foundation for high-speed volumetric cellular-level tomography.

No abstract