Dynamic-parallel MR image reconstruction based on adaptive coil sensitivity estimation

Imaging speed is crucial to dynamic Magnetic Resonance Imaging (dMRI) where multiple images must be obtained in a single examination. To accelerate imaging speed, dMRI can be performed with parallel MRI (pMRI) technique by sparsely sampling the k-space of each time frame. However, pMRI reconstruction requires precise estimation of sensitivity maps of receiving coils, which is difficult to obtain with the existing methods. Imperfect estimation of sensitivity maps can dramatically increase artifacts in the reconstructed image. This paper proposes an adaptive sensitivity estimation method to capture the temporal variation of k-space frames and to estimate the unacquired dataset. The estimated dataset is used to create the full k-space for each frame, from which the time varying sensitivity maps are estimated. Experiments on real dynamic-parallel MRI data demonstrate that the proposed method significantly improves the quality of the reconstructed MR images.