An Optimal Algorithm for Parallel MRI in Presence of Motion Artifacts

Parallel imaging methods provide accelerated multiple coil MR image acquisitions via reconstruction of sub-sampled It- space data. However, the images from component array coil may have strong artifacts due to motion, which in turn have the significant effect on reconstructed composite image. The least squares solution can get arbitrarily wrong when high residual errors are used in the estimation. In this article, these corrupted data are regarded as outliers in observed datum. We present a new algorithm for SENSE which is optimal in presence of artifacts or noise. We apply annealing M-estimator to make solutions insensitive to the influence caused by outliers. Phantom as well as in vivo experiments show significant improvements image quality using the new method. In addition, the benefits are demonstrated compared with conventional SENSE technique.