Quantifying 3-D Vascular Structures in MRA Images Using Hybrid PDE and Geometric Deformable Models
- 4 October 2004
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 23 (10), 1251-1262
- https://doi.org/10.1109/tmi.2004.834612
Abstract
The aim of this paper is to present a hybrid approach to accurate quantification of vascular structures from magnetic resonance angiography (MRA) images using level set methods and deformable geometric models constructed with 3-D Delaunay triangulation. Multiple scale filtering based on the analysis of local intensity structure using the Hessian matrix is used to effectively enhance vessel structures with various diameters. The level set method is then applied to automatically segment vessels enhanced by the filtering with a speed function derived from enhanced MRA images. Since the goal of this paper is to obtain highly accurate vessel borders, suitable for use in fluid flow simulations, in a subsequent step, the vessel surface determined by the level set method is triangulated using 3-D Delaunay triangulation and the resulting surface is used as a parametric deformable model. Energy minimization is then performed within a variational setting with a first-order internal energy; the external energy is derived from 3-D image gradients. Using the proposed method, vessels are accurately segmented from MRA data.This publication has 28 references indexed in Scilit:
- Quantifying vascular structures in MRA images using hybrid PDE and geometric deformable modelsPublished by SPIE-Intl Soc Optical Eng ,2004
- Model-based multiscale detection of 3D vesselsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Anatomical object volumes from deformable B-spline surface modelsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Three-dimensional multi-scale line filter for segmentation and visualization of curvilinear structures in medical imagesMedical Image Analysis, 1998
- A geometric snake model for segmentation of medical imageryIEEE Transactions on Medical Imaging, 1997
- Minimal surfaces based object segmentationIEEE Transactions on Pattern Analysis and Machine Intelligence, 1997
- Volumetric Segmentation of Medical Images by Three-Dimensional BubblesComputer Vision and Image Understanding, 1997
- Geodesic Active ContoursInternational Journal of Computer Vision, 1997
- A multi-scale line filter with automatic scale selection based on the Hessian matrix for medical image segmentationLecture Notes in Computer Science, 1997
- Quantitative coronary angiography with deformable spline modelsIEEE Transactions on Medical Imaging, 1997