Meeting Banner
Abstract #2804

Precomputed Greens Functions for Fast Electromagnetic Simulation with Realistic Human Body Models

Jorge Fernandez Villena1, Amit Hochman2, Luis Miguel Silveira1, Elfar Adalsteinsson2, 3, Lawrence L. Wald, 34, Jacob K. White2, Luca Daniel2

1INESC ID Instituto de Engenharia de Sistemas e Computadores Investigao e Desenvolvimento, Lisbon, Portugal; 2Research Laboratory of Electronics, Massachusetts Institute of Technology, Cambridge, MA, United States; 3Harvard-MIT Division of Health Sciences Technology, Cambridge, MA, United States; 4Department of Radiology, Martinos Center for Biomedical Imaging, Massachusetts General Hospital, Charlestown, MA, United States

We propose a fast electromagnetic (EM) simulation of RF coils operating in the presence of realistic human body models (RHBM), based on frequency-domain integral equation methods. We introduce the RHBM Greens functions (GF), which mimic the effect of the RHBM inside an EM field due to coils currents. This RHBM GF needs to be precomputed only once for a given complex body model, and then it can be combined with traditional Boundary Element Methods that only discretize the coil surface. This scheme generates smaller models that are very fast to solve, allowing for efficient coil design exploration.

Abstract PDF Poster PDF

Keywords

absolute accurately admittance approaches approx around array assembled audience basis becomes biomedical body boundary captured channel circuitry coil coils combination commercial commonly community component components compressed computation computations computed computing conductors configuration confined consuming contributions days dense describe design determined dipole dipoles discretized distributed distribution driving duke electric electromagnetic electronics element elements engineers environments equation error evaluated example excitation expressed extremely fast field fields finally finite free function functions generated generation geometries green grid health hence human important indeed integral intensive involves knowledge known like linear literature magnitude matrices matrix memory model models must needed negligible numerically operating optimization optimize port potential practical precomputed presence principal problems processors proposed quickly radiology realistic relates represent required reusing scattering sciences segments server shared shorter simple simply simulation simulations solution solve solved solver solvers solving source sources space specifically stage stored summing system takes target technology thousands took transmission typically unknowns vectors volume whereas