Dynamic Deformable Models for 3D MRI Heart Segmentation

Zhaosheng Bao, Leonid Zhukov, Igor Guskov, John Wood, David Breen

Proceedings of the International Society for Optical Engineering (SPIE)

Volume 4684, Pages 398-405, 2002

Motivation

Cine magnetic resonance imaging (MRI) is the gold-standard for measuring myocardial mass and function, which is important in diseases such as coronary artery disease, myocarditis, cardiomyopathy, hypertension, diabetes, and cancer. Unfortunately, postprocessing remains laborious and error-prone even though commercial ``computer-assisted" boundary detection algorithms are used. As a result of these difficulties, quantitative function analysis is used only for selected cases even though it is potentially useful for a much larger patient population.

Goal of This Research

The authors wish to develop a clinically useful semi-automatic method to reduce postprocessing of cardiac cine-MRI images.

Goal of This Paper

This paper presents a semi-automatic method that allows a user to quickly extract a 3D model of heart structures. The method essentially starts with an initial triangle mesh specified by the user and allows the mesh to deform within the constraints set by the user. These constraints include features the user selects in addition to mesh properties such as edge length, vertex valence, and aspect ratio which must stay within the ranges the user specifies.

The method uses a recently developed dynamic meshing algorithm that maintains an evenly sampled high-quality mesh throughout the deformation process. The method also uses a new accurate curvature metric for surface meshes to smooth the model during deformation.

Related Work

Results

This work is preliminary. For future work the authors will incorporate an implicit integration scheme to make the solution more stable. The authors anticipate that accurate segmentation will reduce postprocessing time by more than 50%, allowing cardiac function analysis to be more widely used in pediatric cardiovascular disease.

Bibliography