A new 3D segmentation method based on the level set technique is proposed. The main contribution is a robust evolutionary model which requires no fine tuning of parameters. A closed 3D surface propagates from an initial position towards the desired region boundaries through an iterative evolution of a specific 4D implicit function. Information about the regions is involved by estimating, at each iteration, parameters of probability density functions. The method can be applied to different kinds of data, e.g for segmenting anatomical structures in 3D magnetic resonance images and angiography. Experimental results of these two types of data are discussed.
Research Team Aly A. Farag Lab Director farag@cvip.louisville.edu Hossam Hassan Research Assistant msabry@cvip.louisville.edu 
Methods
A level set function is employed for each class region. This approach works with multi-modal images while conventional approaches are suitable only for bi-modal images. Each class region is described by a Gaussian model (pdf function). The speed function of the PDE of each level set depends on these pdfs. The mean, the variance, and the prior probability of each class are estimated iteratively. The level set function evolves starting from a balloon inside the anatomical structure. Our approach overcomes conventional level sets methods problems. It is suitable for multi-modal data (e.g., MRI, MRA, CT).
Results
MRI-Brain Segmentation
MRA-Blood Vessels Segmentation
Publications
A.A. Farag and H. Hassan, Adaptive Segmentation of Multi-modal 3D Data Using Robust Level Set Techniques, Proc. of International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI’04), Saint Malo, France, September 26-29, 2004, pp. 143-150.[PDF]
Acknowledgement / Sponsors
We would like to thank the University of Louisville for its sponsorship.
CVIP Lab 