Biomedical Engineering Reference
In-Depth Information
2
PARAMETRIC CONTOUR MODEL IN
MEDICAL IMAGE SEGMENTATION
Bipul Das
Imaging Technology Division, GE India
Technology Centre, Bangalore, India
Swapna Banerjee
Department of Electronics and ECE, Indian
Institute of Technology, Kharagpur India
The model-based technique offers a unique and efficient approach toward medical image
segmentation and analysis due to its power to unify image information within a physical
framework. Of the model-based techniques, the deformable model is most effectively used
for its ability to unify image statistics — both local and global — in a geometrically con-
strained framework. The geometric constraint imparts a compact formof shape information.
This chapter reviews one of the most promising and highly used deformable approaches:
the active contour model in medical image analysis. The active contour model is one of
the most effective approaches due to its flexibility to adapt to various anatomical shapes
while constraining the local geometric shape constraint. Within the geometric paradigm,
local image statistics and regional information has been effectively used in segmentation
purposes. In addition, various forms of a-priori information can be incorporated into this
model. Active contour models are capable of accommodating a wide range of shape vari-
ability over time and space. The active contour also has to overcome the limitation of
topological adaptibility by introducing a topology adaptive model. This chapter details the
development and evolution of the active contour model with the growing sophistication of
medical images.
