Joint Altered Calibration with Linear Interpolation for Cross-Sampling

Among the methods of sampling trajectories, cross-sampling can provide more autocalibration signal (ACS) data and enables better incoherent sampling with lower gradient control system requirements. However, there were still challenges regarding magnetic field inhomogeneities. So in this study, a self-calibrated off-resonance correction is proposed to eliminate the inconsistency between the orthogonal readout data. This method employs a linear prediction operator to estimate the cross-sampled dat. The corrected data undergo the same reconstruction process as that used in conventional generalized autocalibrating partially parallel acquisitions (GRAPPA), in order to recover the missing k-space data and generate the final image. We will compare this calibration method with the traditional first-order k -space trajectory correction for cross-sampling GRAPPA and the experimental results on phantoms and human subjects demonstrate that the images reconstructed by this method exhibit fewer artifacts and higher signal-to-noise ratio (SNR).