The Liver Segmentation has been perform by
This method use a reference model copute from Visible Man data of the
National Library of Medicine.
Using properties of simplex meshes, reference mesh is deformed using hybrid constraints, local and global. We illustrate here the liver segmentation in abdominal CT-scans. The segmentation requires differents steps.
rigid and affine registration
hybrid deformations (locality factor with a small value)
Déformations hybrides (locality factor with a high value)
Result of this segmentation is then used as a mask to reduce initial image to the Region of Interest.