Fast reduction of undersampling artifacts in radial MR angiography with 3D total variation on graphics hardware

Abstract

Objective-Subsampling of radially encoded MRI acquisitions in combination with sparsity promoting methods opened a door to significantly increased imaging speed, which is crucial for many important clinical applications. In particular, it has been shown recently that total variation (TV) regularization efficiently reduces undersampling artifacts. The drawback of the method is the long reconstruction time which makes it impossible to use in daily clinical practice, especially if the TV optimization problem has to be solved repeatedly to select a proper regularization parameter.Materials and Methods-The goal of this work was to show that for the case of MRAngiography, TV filtering can be performed as a post-processing step, in contrast to the common approach of integrating TV penalties in the image reconstruction process. With this approach it is possible to use TV algorithms with data fidelity terms in image space, which can be implemented very efficiently on graphic processing units (GPUs). The combination of a special radial sampling trajectory and a full 3D formulation of the TV minimization problem is crucial for the effectiveness of the artifact elimination process.Results and Conclusion-The computation times of GPU-TV show that interactive elimination of undersampling artifacts is possible even for large volume data sets, in particular allowing the interactive determination of the regularization parameter. Results from phantom measurements and in vivo angiography data sets show that 3D TV, together with the proposed sampling trajectory, leads to pronounced improvements in image quality. However, while artifact removal was very efficient for angiography data sets in this work, it cannot be expected that the proposed method of TV post-processing will work for arbitrary types of scans.

Calibration-Less Multi-coil MR image reconstruction

“…Radial sampling is by far the best suited practical and efficient sampling scheme for CS based MR image reconstruction techniques [43,44]. It is very fast and is possible to be used for real-time MR data acquisition [45] and is robust to object motion and ghosting artifacts [46].…”

“…Reconstructing images from radial sampling is also more difficult compared to uniform Cartesian sampling, since it cannot be obtained by inverse 2D Fourier transform. In most cases, an optimization problem is solved in order to reconstruct the image [43,44,46]. Radial sampling has also been used for parallel MRI in the past by incorporating TV regularization within the SENSE framework [48].…”

A Guide to the TV Zoo

“…based on wavelets, which can take care of the typical slopes arising between edges in MR images, cf. [32,103,115,[118][119][120].…”

Medical image processing on the GPU – Past, present and future

“…Yang et al (2011) used a GPU to accelerate correction of geometric distortions in echo planar imaging (EPI) data, while Knoll et al (2010) instead used total variation based regularization to reduce undersampling artefacts in radial MR angiography. Deng et al (2011) accelerated the design of multidimensional radio frequency pulses, which can be used to improve high-field imaging.…”