Options
2006
Conference Paper
Titel
Evaluation of active appearance models for cardiac MRI
Abstract
Robust delineation of short-axis cardiac magnetic resonance images (MRI) is a fundamental precondition for functional heart diagnostics. Segmentation of the myocardium and the left ventricular blood pool allows for the analysis of important quantitative parameters. Model-based segmentation methods based on representative image data provide an inherently stable tool for this task. We present an implementation and evaluation of 3-D Active Appearance Models for the segmentation of the left ventricle using actual clinical case images. Models created from varying random data sets have been evaluated and compared with manual segmentations.