Owen Carmichael
Neurology Department, UC Davis
Thursday, October 13, 2005
1131 Kemper Hall
3 :10-4:00 p.m.
Image Processing Methods For Large-Scale Studies of Neurodegenerative Disease
A recent explosion in the quality and amount of biomedical image data has brought ever-smaller biological structures into focus and provided researchers with unprecedented views of the evolution of diseases over time. In this talk I will describe my efforts over the past two years to use computational techniques to intelligently summarize the contents of large biomedical image databases in ways that are too complex and tedious for humans to accomplish. I will focus on attempts to evaluate the accuracy of off-the-shelf, public-domain methods for fully-automated delineation of brain structures in MR images of subjects with Alzheimer's Disease (AD), Mild Cognitive Impairment (MCI), and AIDS. I will then describe our initial findings in applying these techniques to study the time course of brain structure characteristics in AD and MCI. I will conclude with a discussion of important future research directions for biomedical image analysis.
Bio
Owen Carmichael recently started a faculty position in the Neurology Department, UC Davis. He received a PhD in Robotics from Carnegie Mellon University in 2003 and the BS degree with high honors in electrical engineering and computer science from the University of California, Berkeley, in 1997. From 2003 to 2005 he was a postdoctoral scholar in the Radiology Department, University of Pittsburgh. His research interests include medical image analysis, neurodegenerative diseases, computer vision, and 3D modeling. He was a National Science Foundation Graduate Student Fellow from 1997 to 2000. He is a member of the IEEE.