Exploring the Use of Proper Orthogonal Decomposition for Enhancing Blood Flow Images Via Computational Fluid Dynamics

Authors:
Robert Mcgregor;Dominik Szczerba;Martin Siebenthal;Krishnamurthy Muralidhar;Gábor Székely
Affiliations:
Computer Vision Laboratory, Department of Electrical Engineering, ETH, Zürich, Switzerland 8092;Computer Vision Laboratory, Department of Electrical Engineering, ETH, Zürich, Switzerland 8092;Computer Vision Laboratory, Department of Electrical Engineering, ETH, Zürich, Switzerland 8092;Department of Mechanical Engineering, Indian Institute of Technology Kanpur, Kanpur, India 208 016;Computer Vision Laboratory, Department of Electrical Engineering, ETH, Zürich, Switzerland 8092
Venue:
MICCAI '08 Proceedings of the 11th International Conference on Medical Image Computing and Computer-Assisted Intervention, Part II
Year:
2008

Citing 2
Cited 1

Model-Based Initialisation for Segmentation

ECCV '00 Proceedings of the 6th European Conference on Computer Vision-Part II
Inter-subject modelling of liver deformation during radiation therapy

MICCAI'07 Proceedings of the 10th international conference on Medical image computing and computer-assisted intervention - Volume Part I

A Fast Alternative to Computational Fluid Dynamics for High Quality Imaging of Blood Flow

MICCAI '09 Proceedings of the 12th International Conference on Medical Image Computing and Computer-Assisted Intervention: Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

Obtaining high quality patient-specific flow velocity information is not an easy task. Available clinical data are usually poorly resolved and contain a significant amount of noise. We propose a novel approach to integrate computational fluid dynamics with measurement data to overcome this difficulty. By performing a proper orthogonal decomposition of simulated blood flow patterns for a given vascular location with various anatomical configurations it is possible to obtain a basis model for flow reconstruction. This is used to interpolate imaging data intelligently without having to perform a full flow simulation for each individual patient. This work focuses on assessing the feasibility of such a method.