PS Model-Based Dynamic Cardiac MRI with Compressed Sensing

摘要

Real-time cardiac MRI is a challenging topic in MRI field. The partial separability (PS) model has been successfully applied to cardiac MR imaging. However, it is necessary to acquire lots of pre-scanned data to accurately estimate the model parameters before image reconstruction. In order to accelerate the speed of the pre-scanned data acquisition, a new method applying compressive sensing (CS) to the PS model is proposed in this paper, in which the low-rank and sparsity properties of dynamic images were used as priori information for MRI reconstruction. The experiment results show that the proposed method can achieve high resolution dynamic MR imaging and overcome the shortcoming of the conventional PS model.