Meeting Banner
Abstract #1434

Fully Automatic Visualization of 4D Flow Data

Aurelien F. Stalder1, Mehmet Akif Gulsun2, Andreas Greiser3, Marie-Pierre Jolly2

1Siemens AG - Healthcare, Erlangen, Germany; 2Imaging and computer vision, Siemens Corporation - Corporate Technology, Princeton, NJ, United States; 3Siemens AG, Erlangen, Germany

Segmentation, analysis and visualization of 4D Flow data often requires manual interaction and can be complex and time-consuming. In order to overcome the complexity of the processing of such data, a fully automatic approach for visualization of 4D Flow data is presented. Based on the assumption that 4D Flow data can be classified in three kinds of regions: air/lungs, static tissues and vessels/ventricles a data clustering technique is first applied to robustly detect flow regions. Then particle traces are seeded everywhere in the flow region so as to produce fully-automatic 4D visualization of flow data.

Keywords

able according accuracy acquisition adjusted aorta appear application arterial assessment assumption attention automatic automatically background become blood capture circulation clinical clustering clusters colors complex computations context contrast contrasts corporate corporation correct creation dataset datasets defined depend deviation differentiate domain emitter entire estimation everywhere executed expectation faster feature features field flow fully gaining generated gradient healthy hence histogram identical identified identify imperfect implementation include included inspected integrated integration interaction jolly jugulars limited localized long lungs magnitude manual manually marginal mask masking materials maximization mixture model must noisy normalization normalized occurring overcome particle particles path patients patterns plot positioning post processing produce produced pseudo pulmonary quantify random registered remains reproducibility requires resolved restricted running scale secondary seeding segmentation segmented separated shorter sign simple slice slightly solution spatial static streamlining strong successful suggest systole table temporal temporally terms threshold tissues tortuous trace traces vascular velocity venous ventricles vessels vision visualization visually volume volunteer volunteers