Browse Prior Art Database

A Robust Approach to Left Ventricle and Myocardium Segmentation from Temporal Sequence of MR Images

IP.com Disclosure Number: IPCOM000172219D
Original Publication Date: 2008-Jul-25
Included in the Prior Art Database: 2008-Jul-25

Publishing Venue

Siemens

Related People

Other Related People:

Abstract

For the diagnosis of cardiovascular disease, magnetic resonance imaging (MRI) is used which provides time-varying three-dimensional imagery of the heart. Physicians are interested in identifying heart chambers, the myocardium, and measuring changes in ventricular blood volume and wall thickening properties over a cardiac cycle. The left ventricle is of particular interest. For the analyses of these images, a need exists for a method of cardiac segmentation. The methods of cardiac segmentation used at present can be classified into different categories. Some of them are: (1) boundary or edge based segmentation (e.g. level set model for segmentation), (2) region-based method for segmentation (e.g. clustering techniques segmentation), (3) thresholding method and graph cut method for segmentation and morphological watersheds. Furthermore, there are model-based methods where the mathematical model is learnt from a large database based on anatomical knowledge of shape, orientation, size, etc.