Parallel magnetic resonance imaging reconstruction by convex optimization

Magnetic resonance imaging (MRI) scanners implement multiple receiver coils to speed up the scan speed, which is called parallel MRI (pMRI), and the image reconstruction of pMRI involves joint estimation of the image and coil sensitivity functions. In this paper, a constraint on the image function determined by magnetic resonance physics is introduced to formulate the joint image and sensitivity reconstruction into a convex problem. It will lead to a global and optimal solution and accurate and efficient computation for pMRI reconstruction. Performance evaluation is carried out using in-vivo MRI data sets to demonstrate performance of the proposed convex optimization method.