Automatic intensity correction of sensitivity encoding (SENSE) for parallel MR imaging

Abstract

For fast Magnetic Resonance Imaging (MRI), we usually use Sensitivity Encoding (SENSE) which is one of the parallel imaging techniques. It reduces scan time by skipping the phase encoding steps. However, in that process, aliasing effect occurs due to the violation of Nyquist sampling criterion. In SENSE, this aliasing effect can be eliminated by using additional information, sensitivity maps. And in calculating those maps, it uses “sum-of-squares” as denominator image. However, this image does not have uniform intensity because of the non-symmetrical property of RF coil arrays. This artifact affects the sensitivity maps, and eventually also degrades the uniformity of non-aliased full-FOV (Field Of View) image (also known as SENSE reconstructed image). Thus, this study is focused on the correction of the non-uniform intensity of “sum-of-squares” and eventually improving the uniformity of SENSE reconstructed image. The intensity correction is achieved by an appropriate curve fitting to the projection data of the target image. We applied the proposed algorithm to phantom images obtained with 8 channel SENSE.