Generation and initialization of stable 3d mass-spring models for the segmentation of the thyroid cartilage

Authors:
Jana Dornheim;Lars Dornheim;Bernhard Preim;Ilka Hertel;Gero Strauss
Affiliations:
Otto-von-Guericke-Universität, Magdeburg, Germany;Otto-von-Guericke-Universität, Magdeburg, Germany;Otto-von-Guericke-Universität, Magdeburg, Germany;Hals-Nasen-Ohren-Universitätsklinik, Universitätsklinikum Leipzig, Leipzig, Germany;Hals-Nasen-Ohren-Universitätsklinik, Universitätsklinikum Leipzig, Leipzig, Germany
Venue:
DAGM'06 Proceedings of the 28th conference on Pattern Recognition
Year:
2006

Citing 4
Cited 0

Using deformable surfaces to segment 3-D images and infer differential structures

CVGIP: Image Understanding
Surface simplification using quadric error metrics

Proceedings of the 24th annual conference on Computer graphics and interactive techniques
A parametric deformable model to fit unstructured 3D data

Computer Vision and Image Understanding
Automatic segmentation of the left ventricle in 3d SPECT data by registration with a dynamic anatomic model

MICCAI'05 Proceedings of the 8th international conference on Medical Image Computing and Computer-Assisted Intervention - Volume Part I

Quantified Score

Hi-index	0.00

Visualization

Abstract

The preoperative planning of primary tumor resections in the larynx region shall be supported by a 3D visualization of the patient-specific anatomy and pathological situation. This requires a segmentation of the larynx cartilage structures from computed tomography (CT) datasets. In our work, we use 3D Stable Mass-Spring Models (SMSMs) for this segmentation task. Thereto, we create a specific 3D deformable shape model for the segmentation of the thyroid cartilage. A new concept for elastic initialization of this model is presented, allowing the deformable model to adapt specifically to patient-specific shape variations and pathological deformations. We show that using our generation and initialization method, prototypical 3D deformable shape models can be adapted to very different patients without any prior training and prior knowledge about new patients' data.