Seminars/Colloquia/Invited Talks

Seminars

Mikael Rousson

Constrained Surface Evolutions for Prostate and Bladder Segmentation in CT Images

PLACE:	Clark 314
EVENT:	CIS Seminar Series
DATE:	September 13, 2005
TIME:	1:00 - 2:00

Abstract

We propose a Bayesian formulation for coupled surface evolutions and apply it to the segmentation of the prostate and the bladder in CT images. This is of great interest to the radiotherapy treatment process, where an accurate contouring of the prostate and its neighboring organs is needed. A purely data based approach fails, because the prostate boundary is only partially visible. To resolve this issue, we define a Bayesian framework to impose a shape constraint on the prostate, while coupling its extraction with that of the bladder.
We propose a nonlinear statistical shape model for level set segmentation which can be efficiently implemented. Given a set of training shapes of the prostate, we perform a kernel density estimation in the low dimensional subspace spanned by the training shapes. In this way, we are able to combine an accurate model of the statistical shape distribution with efficient optimization in a finite-dimensional subspace. In a Bayesian inference framework, we integrate the nonlinear shape model with a nonparametric intensity model and a set of pose parameters. Constraining the segmentation process makes the extraction of both organs' shapes more stable and more accurate. We present some qualitative and quantitative results on a few data sets, validating the performance of the approach.

Brief biography

Mikael Rousson was born in Annonay, France, in 1978. He graduated from Ecole Superieure en Sciences Informatiques, Sophia-Antipolis, France in 2001. In 2004, he received the PhD degree in computer science and signal processing from University of Nice, Sophia-Antipolis, France. Since November 2004, he has been working as a research scientist at Siemens Corporate Research, Princeton, NJ. His research interests include geometric methods, shape and image statistics.