際際滷

際際滷Share a Scribd company logo
Automatic 4D (3D+time)
Segmentation of Cardiac MRI
CPCC & KLAB
UC Irvine
August 2014
Agenda
 Introduction
 Current Work
 Future work
Introduction
 LV and Myocardium
Segmentation
Challenges in CMR Segmentation
 Poor contrast between myocardium and surrounding
 Diverse brightness in the LV cavity due to blood flow
 Presence of papillary muscles
 Segmentation complexity of apical
and basal slice images
 Inherent noise caused by motion
artifacts, heart dynamics
 Shape and intensity variability of the LV
3D Automatic Segmentation
 Developed Algorithm
Phase 0: Automatic Localization
 Input: Raw Image Output: Chamber Location
Automatic
Chamber Detection
Phase I: 3D Segmentation of LV
3D
Segmentation
Phase II: Intra-Chamber Inclusion
Intra-Chamber
Inclusion
Phase III: Myocardial Segmentation
3D
Segmentation
Future Work: 4D (3D+time) Segmentation
Moving Heart

More Related Content

Automatic 4D (3D+time) Segmentation of Cardiac MRI