3D object segmentation using deformable models
Date of Issue2007
School of Electrical and Electronic Engineering
This thesis presents research work on deformable surface model for 3D object segmentation. Over the past decades, there have been many research activities in 3D object segmentation using 3D and 2D deformable models. Full 3D methods will produce much better results than those obtained based on the 2D ones. Contextual intensity information of one voxel in one direction will be lost in each 2D image. Furthermore, a post-processing step is required to connect the sequence of 2D contours in a continuous surface. Reconstruction of the surface will be difficult if the topology of 2D contours is complicated. Therefore, it is desired to detect objects in 3D space directly to avoid the shortcomings of 2D methods, especially in application to 3D medical image analysis.
DRNTU::Engineering::Electrical and electronic engineering::Electronic systems::biometrics