Regularized motion correction algorithm for diffusion weighted magnetic resonance imaging
FANG Sheng1, ZHAO Li1, YING Kui, CHENG Jianping1(Key Laboratory of Particle and Radiation Imaging of the Ministry of Education, Department of Physical Engineering, Tsinghua University, Beijing 100084, China)
Motion artifacts in multi-shot diffusion weighed imaging are removed by a motion correction algorithm that converges faster and does not require a high resolution navigator image. This method improves the conjugate gradient algorithm using Tikhonov regularization. With a navigator image as prior knowledge, the method can effectively exploit the navigator data, so that residual artifacts caused by a low resolution navigator image are reduced. The method accelerates convergence of the conjugate gradient method using parallel imaging to reduce the scan time. Both simulations and in vivo experiments demonstrate that the method effectively reduces motion-induced artifacts and the mean square error of the reconstruction, which is important for improved diagnosis.