An iterative self-consistency parallel imaging reconstruction method based on transformation learning and joint sparsity

Abstract

The invention relates to an iterative self-consistency parallel imaging reconstruction method based on transformation learning and joint sparsity, and belongs to the technical field of medical magnetic resonance imaging. Based on an iteration self-consistency parallel imaging reconstruction problem, the invention provides a Cartesian iteration self-consistency parallel magnetic resonance imaging reconstruction method combining transformation learning and sparse regularization terms. and carrying out solution by using a variable separation (VS) technology and an alternating direction multipliermethod (ADMM) technology. The method comprises the following steps of: carrying out solution by using an ADMM method, Simulation experiments on the two actual data sets show that compared with othercomparison methods, the new algorithm provided by the invention can obtain better reconstruction quality.

Description

technical field

[0001] The invention relates to an iterative self-consistent parallel imaging reconstruction method based on transformation learning and joint sparsity, belonging to the technical field of medical magnetic resonance imaging. Background technique

[0002] Parallel magnetic resonance (Magnetic Resonance, MR) imaging is a well-known accelerated imaging method. Its advantage is to reduce the sampling time of MR imaging by receiving the spatial sensitivity information through the multi-coil array receiver. In the past two decades, many parallel imaging reconstruction methods have been proposed, and these methods are different due to different ways of using the sensitivity information; for example, the Sensitivity Encoding (SENSitivity Encoding, SENSE) method uses explicit sensitivity information to perform Refactored. The main limitation of this method in practical application is that it is difficult to measure the sensitivity information accurately. The other ...

Images