A Practical stopping Criterion for Iterative Non-Cartesian SENSE Reconstruction

Abstract

In non-Cartesian SENSE reconstruction based on conjugate gradient (CG) iteration method, the iteration very often exhibits a ¡ semi-convergence¡± behavior which can be characterized as initial convergence towards the exact solution and later divergence. This phenomenon makes a suitable stopping criterion very critical for automatic implementation of this reconstruction strategy. In this study a practical stopping rule for iterative SENSE reconstruction is proposed. By monitoring the residual norm along the iteration, this stopping criterion can automatically stop the process where artifact and SNR are well compromised and good overall image quality are achieved.