Meeting Banner
Abstract #1246

Evaluation of Divergence-Free Correction Algorithms in High Resolution 4-D Flow Images of Cranial Vasculature.

Michael Loecher1, Steven Kecskemeti, Kevin M. Johnson1, Patrick Turski2, Oliver Wieben1, 2

1Medical Physics, University of Wisconsin Madison, Madison, WI, United States; 2Radiology, University of Wisconsin Madison, Madison, WI, United States

This study investigates the performance of a finite difference method and a radial basis function algorithm for removing divergence from 4D-flow images. Noise reduction and visualization improvements are compared in a phantom and in vivo by comparing RMSE, noise levels, and streamline lengths. Overall the RBF method performs better, generally reducing noise by around 25% and improving streamline length by around 22%.

Keywords

acquisition aorta apex applied arteries artery assistance assume ates basis better blood broad calculating calculations carotid carotids clinical comp conditions considered consistent contain coronal corr correction coupled cover create dataset datasets decrease defined derivatives deviation dial direct display divergence eduction equation error evaluation explained extend fast features field fields finite fitting flirt flow free function functions funding gains generated gradient gratefully hows identity important improvements in vivo included includes initial internal interpolation inverse knowledge larger length lengths made maps matrix measured measures medium motion mount mutually noise norm note oliver option outside overall oxford pair pairs phantom physics plying prevent previous projection projections radial radially radiology reasonable reasons reconstructed reduce reduction removal remove represent represents required resolution resolutions routine scanner sharing short significantly slice slightly solve solved solves spatial straight stream streamline streamlines strong subsequent subtracting takes taking temporal transform true tube utilize vascular velocities velocity vessel viable view visualization volume volunteer wing