Orientation histograms as shape priors for left ventricle segmentation using graph cuts

Authors:
Dwarikanath Mahapatra;Ying Sun
Affiliations:
Department of Electrical and Computer Engineering, National University of Singapore, Singapore;Department of Electrical and Computer Engineering, National University of Singapore, Singapore
Venue:
MICCAI'11 Proceedings of the 14th international conference on Medical image computing and computer-assisted intervention - Volume Part III
Year:
2011

Citing 7
Cited 0

Fast Approximate Energy Minimization via Graph Cuts

IEEE Transactions on Pattern Analysis and Machine Intelligence
Diffusion Snakes: Introducing Statistical Shape Knowledge into the Mumford-Shah Functional

International Journal of Computer Vision
Interactive Graph Cut Based Segmentation with Shape Priors

CVPR '05 Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05) - Volume 1 - Volume 01
Star Shape Prior for Graph-Cut Image Segmentation

ECCV '08 Proceedings of the 10th European Conference on Computer Vision: Part III
Graph-based knowledge-driven discrete segmentation of the left ventricle

ISBI'09 Proceedings of the Sixth IEEE international conference on Symposium on Biomedical Imaging: From Nano to Macro
Graph cuts framework for kidney segmentation with prior shape constraints

MICCAI'07 Proceedings of the 10th international conference on Medical image computing and computer-assisted intervention - Volume Part I
Segmenting and tracking the left ventricle by learning the dynamics in cardiac images

IPMI'05 Proceedings of the 19th international conference on Information Processing in Medical Imaging

Quantified Score

Hi-index	0.00

Visualization

Abstract

Poor contrast in magnetic resonance images makes cardiac left ventricle (LV) segmentation a very challenging task. We propose a novel graph cut framework using shape priors for segmentation of the LV from dynamic cardiac perfusion images. The shape prior information is obtained from a single image clearly showing the LV. The shape penalty is assigned based on the orientation angles between a pixel and all edge points of the prior shape. We observe that the orientation angles have distinctly different distributions for points inside and outside the LV. To account for shape change due to deformations, pixels near the boundary of the prior shape are allowed to change their labels by appropriate formulation of the penalty and smoothness terms. Experimental results on real patient datasets show our method's superior performance compared to two similar methods.