1: \begin{abstract}

2: Magnetic Resonance Imaging (MRI) is a kind of medical imaging technology used for diagnostic imaging of

3: diseases, but its image quality may be suffered by the long acquisition time.

4: The compressive sensing (CS) based

5: strategy may decrease the reconstruction time greatly, but it needs efficient reconstruction algorithms

6: to produce high-quality and reliable images.

7: This paper focuses on

8: the algorithmic improvement for the sparse reconstruction of CS-MRI, especially considering

9: a non-smooth convex minimization problem which is composed of the sum of a total variation regularization term and a $\ell_1$-norm term of the wavelet transformation.

10: The partly motivation of targeting the dual problem is that the dual variables are involved in relatively low-dimensional subspace.

11: Instead of solving the primal model as usual, we turn our attention to its associated dual model composed of

12: three variable blocks and two separable non-smooth function blocks.

13: However, the directly extended alternating direction method of multipliers (ADMM)

14: must be avoided because it may be divergent, although it usually performs well numerically.

15: In order to solve the problem, we employ a symmetric Gauss-Seidel (sGS) technique based ADMM.

16: Compared with the directly extended ADMM, this method only needs one additional iteration, but its convergence can be guaranteed theoretically.

17: Besides,  we also propose a generalized variant of ADMM because this method has been illustrated to be efficient for solving

18: semidefinite programming in the past few years.

19: Finally, we do extensive experiments on MRI reconstruction using some simulated and real MRI images

20: under different sampling patterns and ratios.

21: The numerical results demonstrate that the proposed algorithms significantly

22: achieve high reconstruction accuracies with fast computational speed.

23: \end{abstract}

24: