Swarthmore College Department of Computer Science

Talk by Joshua Stough, University of North Carolina, Chapel Hill

Image appearance models for automatic medical image segmentation
Friday, February 29, 2008
4 pm in the main cs lab, Room 240
Cookies at 3:45!

My research interests are computer vision in general and automatic medical image segmentation in particular. Medical image segmentation refers to the task of delineating objects such as organs in 2D and 3D medical images. Quality segmentation is a critical aspect of many medical applications, including radiation therapy for the treatment of cancer, surgical planning, and diagnosis. I focus on one aspect of automatic segmentation methods, the appearance model. An effective appearance model provides a "fingerprint" for an object, describing the image intensity patterns in and around the object of interest. Using such appearance models improves automatic segmentation results. Please see http://cs.unc.edu/~stough/research.html for more.