A learning framework for the automatic and accurate segmentation of cardiac tagged MRI images

Authors:
Zhen Qian;Dimitris N. Metaxas;Leon Axel
Affiliations:
Center for Computational Biomedicine Imaging and Modeling (CBIM), Rutgers University, New Brunswick, New Jersey;Center for Computational Biomedicine Imaging and Modeling (CBIM), Rutgers University, New Brunswick, New Jersey;Department of Radiology, New York University, New York
Venue:
CVBIA'05 Proceedings of the First international conference on Computer Vision for Biomedical Image Applications
Year:
2005

Citing 4
Cited 3

Active shape models—their training and application

Computer Vision and Image Understanding
Automated Segmentation of the Left and Right Ventricles in 4D Cardiac SPAMM Images

MICCAI '02 Proceedings of the 5th International Conference on Medical Image Computing and Computer-Assisted Intervention-Part I
A decision-theoretic generalization of on-line learning and an application to boosting

EuroCOLT '95 Proceedings of the Second European Conference on Computational Learning Theory
Face alignment using statistical models and wavelet features

CVPR'03 Proceedings of the 2003 IEEE computer society conference on Computer vision and pattern recognition

Cardiac Motion Recovery and Boundary Conditions Estimation by Coupling an Electromechanical Model and Cine-MRI Data

FIMH '09 Proceedings of the 5th International Conference on Functional Imaging and Modeling of the Heart
Left Ventricle Segmentation Using Diffusion Wavelets and Boosting

MICCAI '09 Proceedings of the 12th International Conference on Medical Image Computing and Computer-Assisted Intervention: Part II
Boosting and nonparametric based tracking of tagged MRI cardiac boundaries

MICCAI'06 Proceedings of the 9th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper we present a fully automatic and accurate segmentation framework for 2D tagged cardiac MR images. This scheme consists of three learning methods: a) an active shape model is implemented to model the heart shape variations, b) an Adaboost learning method is applied to learn confidence-rated boundary criterions from the local appearance features at each landmark point on the shape model, and c) an Adaboost detection technique is used to initialize the segmentation. The set of boundary statistics learned by Adaboost is the weighted combination of all the useful appearance features, and results in more reliable and accurate image forces compared to using only edge or region information. Our experimental results show that given similar imaging techniques, our method can achieve a highly accurate performance without any human interaction.