Endocardial 3D Ultrasound Segmentation using Autocontext Random Forests

Please use this identifier to cite or link to this publication: http://hdl.handle.net/10380/3484
In this paper, we present the use of a generic image segmentation method, namely a succession of Random Forest classifiers in an autocontext framework, for the MICCAI 2014 Challenge on Endocardial 3D Ultrasound Segmentation (CETUS). The proposed method segments each frame independently in 90 sec, without requiring temporal information such as end-diastolic or end-systolic time points nor any registration. For better segmentation accuracy, non-local means denoising can be applied to the images at the cost of an increased run-time. The mean Dice score on the testing dataset was 84.4% without denoising and 86.4% with denoising. The originality of our approach lies in the introduction of two classes, the myocardium and the mitral valve, in addition to the left ventricle and the background classes, in order to gain contextual information for the segmentation task.

There is no review at this time. Be the first to review this publication!

Quick Comments

Download All
Download Paper , View Paper

Statistics more
Global rating: starstarstarstarstar
Review rating: starstarstarstarstar [review]
Paper Quality: plus minus

Information more
Categories: Segmentation, Unsupervised learning and clustering
Keywords: Segmentation, 3D ultrasound imaging
Export citation:


Linked Publications more
InsightToolkit Kinetic Analysis (itk::ka) Library InsightToolkit Kinetic Analysis (itk::ka) Library
by Dowson N., Baker C., Raffelt D., Smith J., Thomas P., Salvado O., Rose S.
Segmentation of Multi-Center 3D Left Ventricular Echocardiograms by Active Appearance Models Segmentation of Multi-Center 3D Left Ventricular Echocardiograms by Active Appearance Models
by Van Stralen M., Haak A., Leung K.E., Van Burken G., Bosch J.G.

View license
Loading license...

Send a message to the author
Powered by Midas