Image-Based Analysis of the Topology and Morphology
of Diseased Lungs
Department of Computer Science
Tuesday, April 10
1111 Biotech Center
Understanding how diseases affect the topology and morphology of human lungs can help medical practitioners develop and apply effective treatments for such diseases. Image-based analysis is emerging as an essential tool to help answer many of the open questions in this area. In this talk, I will focus on problems in image registration and computational modeling of the lungs. I will first present the problem of aligning an image of healthy lungs to an image of the same lungs inflicted with acute respiratory distress syndrome. Conventional non-rigid registration approaches do not correctly align all of the visible anatomy. One reason is that the similarity measures used to achieve alignment are assumed to apply equally well to all areas of the images to be registered. I will present work on using similarity measures non-uniformly during an image registration process and show how thisimproves the resulting alignment. I will also present the problem of determining which airways cause degradation in the lung function of asthmatics. Generic models of the airway tree have been used to reveal that small airways are the significant cause of lung degradation. I will
show an approach for generating personalized models of the human airway tree using Hyperpolarized Helium MR and Cryosection images. These personalized models significantly differ from a generic model and may allow for a more refined characterization of which airways cause degradation of lung function.
Bio: William Mullally is currently a Ph.D. candidate in Computer
Science at Boston University where he is a member of the Image and Video Computing Group. His research interests are medical image analysis and its applications in medicine, biomedical engineering, and bioinformatics.
Airway Tree on He MRI
Return to seminar list