Exploiting typical clinical imaging constraints for 3D outer bone surface segmentation

We present a method for extracting outer bone surfaces from a 3D CT (computer tomography) image sequence using a novel segmentation scheme on each image. A 3D mesh of the bone surface is then generated using the marching cubes algorithm. The new segmentation algorithm makes use of several imaging co...

Full description

Saved in:
Bibliographic Details
Main Authors: Mack, C., Mogallapu, V., Willis, A., Weldon, T.P.
Format: Conference Proceeding
Language:English
Subjects:
Biomedical imaging
Bones
Computed tomography
Data mining
Humans
Image segmentation
Medical diagnostic imaging
Orthopedic surgery
Shape
Surface treatment
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a method for extracting outer bone surfaces from a 3D CT (computer tomography) image sequence using a novel segmentation scheme on each image. A 3D mesh of the bone surface is then generated using the marching cubes algorithm. The new segmentation algorithm makes use of several imaging constraints which greatly simplify the problem including : (i) the cross-sectional size of a bone is approximately known and (ii) the geometric shape of a cross-section is approximately known. In clinical practice using commercial CT scanners, these quantities are typically known and serve to greatly simplify the segmentation problem. By segmenting the image data, the algorithm is capable of uniquely extracting the bone outer surface in contrast to other methods which often include extra surfaces or surfaces with holes. This paper presents the segmentation method and shows results for extracting tibia bone outer surfaces.
ISSN:1091-0050
1558-058X
DOI:10.1109/SECON.2007.342956