Meeting Banner
Abstract #4192

Minimum Squared Error Estimate of Electrical Properties from B1 Maps

Selaka Bandara Bulumulla1, Seung-Kyun Lee1, Ileana Hancu1

1GE Global Research, Niskayuna, NY, United States

In pixel by pixel calculation of conductivity and permittivity from B1 maps, the additive noise in B1 can impair the quality of reconstructed electrical properties images. In previous work, discarding of non-physical values (e.g. negative conductivity) or filtering of B1 maps have been considered. Discarding values lead to missing pixels in the electrical properties images and filtering may remove spatial variation containing information on electrical properties. In this work, we consider minimum squared error estimate of conductivity and permittivity for a region of constant electrical properties and show promising results for a phantom with conductivity contrast.

Keywords

absorption added addressed agreed applications audience avoid axial better bottom boundaries boundary breast calculation coil compartment compartments composite concentrations concepts conditions conductivity consider constant copper cost define derivative determination development diameter dielectric discarding distilled electric electrical equation equations error errors essential experiment extending extracting feasibility field fields filtering find fluids frequency function global guide half head hence ideal identified identify impaired in vivo inside knowledge lead local magnitude manually mapping maps material materials measured minimize missing negative noise omitted outer part partial pathology personnel phantom physical pixel pixels planning potential prepared previous probe promising properties proposed quadratic quality radio reconstructed reconstruction remove robust satisfy segmentation setting smoothing spatial sphere spheres spin squared sulfate table target tissue transitions transmit treatment types underlying validate variations various water zero