Title:
Representing Tubular Objects in Medical Images:
Visualizing Intracranial Vessels for Surgical Planning

Abstract:
This talk will focus on a novel technique for generating representations
of tubular objects in 3D medical data. Tubular objects can be
characterized as smoothly varying, yet possibly branching, structures in
3D that have nearly circular cross sections.  Tubular objects are
abundant in medical images, e.g., vessels, bones, ducts, spinal cords,
and bowels.

While other techniques have been suggested for segmenting tubular
objects, the method presented rapidly generates accurate and consistent
tubular representations with minimal user interaction by exploiting the
geometry of tubes. Specifically, tubular objects defined via contrast
are special in that blurring produces a central intensity ridge that
well approximates the objects' central skeleton. Our method operates by
traversing those central skeletons. Once extracted, those central
skeletons also serve to stabilize a width estimation process.

This talk will illustrate the use of our method in the formation and
manipulation of vascular and bronchial trees for surgical planning.

Here's my general CV:
R2 Technologies, Inc., Los Altos, CA
January, 1998 - Present
	Consultant.   Performing applied research in the area of 
computer-aided diagnosis.

University of North Carolina at Chapel Hill
February, 1997 - Present
	Research Assistant Professor of Radiology and Adjunct Assistant
Professor of Computer Science.   Performing applied research regarding
(1) computer-aided diagnosis of direct-digital mammograms using mixture
modeling and multi-scale measures of intensity and texture, (2)
visualizing vessels for pre-operative planning and intra-operative
guidance, and (3) statistical methods for associating the shape of
structures in the brain with brain diseases.

McDonnell Aircraft Company, St. Louis, MO
October, 1989 - December, 1992
	Senior Engineer, Technical Leader, Neural Networks Laboratory.   
Group focused on the practical application of neural networks to avionic
products and processes.   Performed basic and applied research in neural
networks and investigated the use of genetic algorithms, fuzzy logic,
and hybrid systems.


=======================================================================
Stephen R. Aylward			 http://www.cs.unc.edu/~aylward
Research Assistant Professor of Radiology		aylward@unc.edu
Adjunct Assistant Professor of Computer Science		 (919) 966-9695

.