Date: May 14, 2009 Time: 10.30am Location: MI 02.13.010

Abstract:Recently, tools of variational calculus have found extensive applications in computer vision. In this talk, I will present variational approaches to medical vision problems in image segmentation and registration. After showing various deformable models and their applications in medical image segmentation, I will talk about coupled partial differential equations that we proposed to estimate vector fields and surfaces to solve a simultaneous segmentation and registration problem. We also explored the utility of inequality constraints applied to estimation of deformation fields in non-rigid registration and tracking. In order to solve inequality constrained vector field estimation problems, we apply tools from the Kuhn- Tucker theorem in optimization theory. Finally, I will present some clinically relevant applications from our current research projects in medical image analysis.

Biography: Gozde Unal received her Ph.D. degree in electrical and computer engineering from North Carolina State University, Raleigh, NC, USA, in 2002, and then was a postdoctoral fellow at the Georgia Institute of Technology, Atlanta, GA, USA for a year. Between 2003 and 2007, she worked as a research scientist at Siemens Corporate Research, Princeton, NJ, USA. She joined the faculty of Sabanci University, Istanbul, Turkey in Fall 2007, where she is currently an assistant professor. Her current research is focused on medical image analysis, segmentation, registration, and shape analysis techniques with applications to clinically relevant problems in MR, CT, US, and intravascular images. She is a Senior Member of the IEEE, and an Associate Editor for the IEEE Transactions on Information Technology on Biomedicine. She is the initiating co-organizer of the biennial CVII-MICCAI Workshop series (Computer Vision for Intravascular Imaging), which were held in 2006 in Copenhagen, Denmark, and in 2008, in New York, USA. (Take a look at Gozde Unal's website)