Accelerating SENSE using compressed sensing

Abstract

Both parallel MRI and compressed sensing (CS) are emerging techniques to accelerate conventional MRI by reducing the number of acquired data. The combination of parallel MRI and CS for further acceleration is of great interest. In this paper, we propose a novel method to combine sensitivity encoding (SENSE), one of the standard methods for parallel MRI, and compressed sensing for rapid MR imaging (SparseMRI), a recently proposed method for applying CS in MR imaging with Cartesian trajectories. The proposed method, named CS-SENSE, sequentially reconstructs a set of aliased reducedfield-of-view images in each channel using SparseMRI and then reconstructs the final image from the aliased images using Cartesian SENSE. The results from simulations and phantom and in vivo experiments demonstrate that CS-SENSE can achieve a reduction factor higher than those achieved by SparseMRI and SENSE individually and outperform the existing method that combines parallel MRI and CS. Magn Reson Med 62:1574 -1584, 2009.

A rapid and robust numerical algorithm for sensitivity encoding with sparsity constraints: Self‐feeding sparse SENSE

“…; generated by alternately solving Eqs. [5] and [7] from any starting point (I 0 ,Î 0 ) converges to (I a ,Î a ), a solution of [4], and the convergence rate is linear, that is,…”

“…For both conventional SENSE (1) and prior information regularized SENSE (26), the noise decorrelation is a routine step. When a Cartesian trajectory is used, model [4] can be further improved by introducing spatially adaptive weights to the regularization term. As the g-factor map (1) is available from the initial I-subproblem, this noise distribution information can be used to improve the effectiveness of Î-subproblem.…”

“…For example, acceleration factors are no more than 3 in most clinical exams for the widely available eight-channel head coil. To reduce the noise and artifact levels, techniques of enforcing sparsity in PPI reconstruction have recently been proposed (3)(4)(5)(6), where ''sparsity'' means that the to-be-reconstructed image has a sparse representation in a known and fixed mathematical transform domain (7). These techniques enforce the sparsities of the to-be-reconstructed image in finite difference domain and wavelet transform domain by minimizing its total variation (TV) norm and the L 1 norm of its wavelet transform.…”

“…(3), impressive reconstructions were presented with acceleration factor as high as 8 in 2D imaging using radial acquisition and a 12-channel head coil. Using a piece-wise constant phantom and vary-density random acquisition scheme, an acceleration factor of 8 was achieved with an eight-channel torso coil (4). Using 2D acceleration and Poisson-disc sampling, an acceleration factor of 5 (2.5 Â 2) was achieved for 3D with a four-channel extremity coil (5).…”

Accelerated Dynamic Cardiac MRI Using Motion-Guided Compressed Sensing with Regional Sparsity

“…We also observed from some of the patients studies that at high acceleration rates, the strong fat signals from the chest and/or the back regions were more overlapped onto the heart region. The CS regularization power was increased for those studies and sometimes moderate residual artifacts 75 could still be observed. In the future, techniques such as fat-saturation RF pulses or outer-volume suppression preparation can be adopted to decrease the influence of the strong fat signals from regions that are not of interest.…”