Quality-based UnwRap of SUbdivided Large Arrays (URSULA) for high-resolution MRI data

In Magnetic Resonance Imaging, mapping of the static magnetic field and the magnetic susceptibility is based on multidimensional phase measurements. Phase data are ambiguous and have to be unwrapped to their true range in order to exhibit a correct representation of underlying features. High-resolution imaging at ultra-high fields, where susceptibility and phase contrast are natural tools, can generate large datasets, which tend to dramatically increase computing time demands for spatial unwrapping algorithms. This article describes a novel method, URSULA, which introduces an artificial volume compartmentalisation that allows large-scale unwrapping problems to be broken down, making URSULA ideally suited for computational parallelisation. In the presented study, URSULA is illustrated with a quality-guided unwrapping approach. Validation is performed on numerical data and an application on a high-resolution measurement, at the clinical field strength of 3T is demonstrated. In conclusion, URSULA allows for a reduction of the problem size, a substantial speed-up and for handling large data sets without sacrificing the overall accuracy of the resulting phase information.