KMID : 1137820070280060791
|
|
ÀÇ°øÇÐȸÁö 2007 Volume.28 No. 6 p.791 ~ p.797
|
|
Developments of Semi-Automatic Vertebra Bone Segmentation Tool using Valley Tracking Deformable Model
|
|
Kim Yie-Bin
Kim Dong-Sung
|
|
Abstract
|
|
|
This paper proposes a semiautomatic vertebra segmentation method that overcomes limitations of both manual segmentation requiring tedious user interactions and fully automatic segmentation that is sensitive to initial conditions. The proposed method extracts fence surfaces between vertebrae, and segments a vertebra using fence-limited region growing. A fence surface is generated by a deformable model utilizing valley information in a valley emphasized Gaussian image. Fence-limited region growing segments a vertebra using gray value homogeneity and fence surfaces acting as barriers. The proposed method has been applied to ten patient data sets, and produced promising results accurately and efficiently with minimal user interaction.
|
|
KEYWORD
|
|
vertebra segmentation, deformable model, valley emphasized image, fence-limited SRG
|
|
FullTexts / Linksout information
|
|
|
|
Listed journal information
|
|
|