B₁^- non-uniformity correction of phased-array coils without measuring coil sensitivity

Parallel imaging can be used to increase SNR and shorten acquisition times, albeit, at the cost of image non-uniformity. B₁^- non-uniformity correction techniques are confounded by signal that varies not only due to coil induced B₁^- sensitivity variation, but also the object's own intrinsic signal. Herein, we propose a method that makes minimal assumptions and uses only the coil images themselves to produce a single combined B₁^- non-uniformity-corrected complex image with the highest available SNR. A novel background noise classifier is used to select voxels of sufficient quality to avoid the need for regularization. Unique properties of the magnitude and phase were used to reduce the B₁^- sensitivity to two joint additive models for estimation of the B₁^- inhomogeneity. The complementary corruption of the imaged object across the coil images is used to abate individual coil correction imperfections. Results are presented from two anatomical cases: (a) an abdominal image that is challenging in both extreme B₁^- sensitivity and intrinsic tissue signal variation, and (b) a brain image with moderate B₁^- sensitivity and intrinsic tissue signal variation. A new relative Signal-to-Noise Ratio (rSNR) quality metric is proposed to evaluate the performance of the proposed method and the RF receiving coil array. The proposed method has been shown to be robust to imaged objects with widely inhomogeneous intrinsic signal, and resilient to poorly performing coil elements.