Modified Hough transform for left ventricle myocardium segmentation in 3-D echocardiogram images
- 1 May 2012
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 19457928,p. 290-293
Abstract
Segmentation of the left ventricular myocardium from 3D echocardiograms is complicated by speckle, artifact, and complex anatomy. Typical segmentation methods benefit from accurate initialization. We propose a two-step Hough transform to find an annular approximation of the myocardium in short-axis echo slices. The method was compared to manual segmentation of 5641 slices by center points and endocardial and epicardial radii. Centers deviated by a mean 3.31 mm, and radii by 2.57 mm and 3.04 mm, respectively. Dice's coefficient of similarity was 0.70. Similarity is higher when apical slices are ignored. 93.8% of all slices and 96.8% of non-apical slices met criteria that suggest they would be appropriate to initialize further segmentation.Keywords
This publication has 7 references indexed in Scilit:
- The evaluation of single-view and multi-view fusion 3D echocardiography using image-driven segmentation and trackingMedical Image Analysis, 2011
- A coupled deformable model for tracking myocardial borders from real-time echocardiography using an incompressibility constraintMedical Image Analysis, 2010
- Statistical shape models for 3D medical image segmentation: A reviewMedical Image Analysis, 2009
- Automatic Detection of Left Ventricle in Echocardiographic Sequences using Radial Search and Temporal Smoothing2007
- LV Motion Tracking from 3D Echocardiography Using Textural and Structural Information2007
- Ultrasound image segmentation: a surveyIEEE Transactions on Medical Imaging, 2006
- Generalizing the Hough transform to detect arbitrary shapesPattern Recognition, 1981