Automatic heart isolation in 3d CT images

Authors:
Hua Zhong;Yefeng Zheng;Gareth Funka-Lea;Fernando Vega-Higuera
Affiliations:
Imaging and Computer Vision, Siemens Corporate Technology, Princeton, NJ;Imaging and Computer Vision, Siemens Corporate Technology, Princeton, NJ;Imaging and Computer Vision, Siemens Corporate Technology, Princeton, NJ;Healthcare Sector, Siemens AG, Forchheim, Germany
Venue:
MCV'12 Proceedings of the Second international conference on Medical Computer Vision: recognition techniques and applications in medical imaging
Year:
2012

Citing 5
Cited 0

Active shape models—their training and application

Computer Vision and Image Understanding
Curve and surface smoothing without shrinkage

ICCV '95 Proceedings of the Fifth International Conference on Computer Vision
Probabilistic Boosting-Tree: Learning Discriminative Models for Classification, Recognition, and Clustering

ICCV '05 Proceedings of the Tenth IEEE International Conference on Computer Vision - Volume 2
Using anatomical knowledge expressed as fuzzy constraints to segment the heart in CT images

Pattern Recognition
Automatic aorta segmentation and valve landmark detection in C-Arm CT: application to aortic valve implantation

MICCAI'10 Proceedings of the 13th international conference on Medical image computing and computer-assisted intervention: Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this chapter, we present an automatic heart segmentation algorithm for the diagnosis of coronary artery diseases (CAD). The goal is to visualize the heart from a cardiac CT image with irrelevant tissues such as the lungs, rib cage, pulmonary veins, pulmonary arteries and left atrial appendage hidden so that doctors can clearly see the major coronary artery trees, aorta and bypass arteries if they exist. The algorithm combines a model-based detection framework with data-driven post-refinements to create a mask for a given cardiac CT image that contains only the relevant part of the heart. The marginal space learning [1] technique is used to localize mesh model or landmark points of different cardiovascular structures in the CT volume. Guided by such detected models, local data-driven voxel-based refinements are employed to produce precise boundaries of the heart mask. The algorithm is fully automatic and can process a 3D cardiac CT volume within a few seconds.