A fast method for computing and correcting intensity inhomogeneities in MRI

In Magnetic Resonance Imaging (MRI) a given tissue may have quite different intensities depending on its location in the image volume. Such intensity inhomogeneities often arise from what is usually called the bias field, and are due to radio frequency inhomogeneities, variations in the static magnetic field and the distance of the tissue to the receiving coil. We present an efficient approach to determine such inhomogeneities from a series of calibration scans, which take less than 40 seconds to acquire. This enables us to efficiently correct both T1 and T2 weighted images. We present the theoretical background, as well as typical results for phantom, liver, pelvic and spine images. ©2008 IEEE