Paul Yushkevich

biography

My doctoral research focused primarily on developing techniques for representing the morphological variability of anatomical structures such as the hippocampus and for automatically segmenting these structures in medical images. Building on the medial representation (m-rep) approach invented by Dr. Pizer, developed a new class of continuous medial representations (cm-rep), which turned out to offer unique advantages over other popular object representations because they allows the space inside of objects to be parameterized by a curvilinear, shape-based coordinate system. After receiving my Ph.D., I spent a year working on a software engineering project funded by the National Library of Medicine (NLM). I led the integration of the SNAP tool, developed at UNC under direction of Dr. Guido Gerig with the Insight Toolkit, a computational platform spearheaded by the NIH/NLM. In the process, the tool was completely redesigned and many new features were added to it. ITK-SNAP was intended to help fill the divide that often exists between scientists who develop new quantitative methods and investigators involved in patient-oriented research. By investing effort in software design, we made the complicated mathematics of level set segmentation easily accessible to users without mathematical expertise. ITK-SNAP has been adopted as a segmentation tool of choice at labs at UNC, Penn, Duke, Yale and other universities and hospitals. My transition from Computer Science to Radiology Research at Penn exposed me to a broad range of new biomedical problems to which the methods that I had been developing could be applied. Working as a research fellow in Dr. Gee's lab, I expanded my expertise beyond object representation and automatic segmentation, publishing papers on whole brain morphometry, normalization and analysis of diffusion tensor MRI, and volumetric reconstruction of the murine brain from histological data. I got my present position in 2006, and I am actively seeking collaborations and funding sources for expanding the research program on geometrical and statistical methods for brain morphometry.

active grants

K25-AG027785-01 (PI)
04/01/07-01/31/12
NIH/NIA
Computational methods for regional hippocampal morphometry in AD

R21 NS061111-01 (PI)
12/01/07-11/30/09
NIH/NINDS
Object-Centric Computational Model for Imaging Analysis

R03 EB008200 (PI)
10/01/07-09/30/08
NIH/NIBIB
Accessibility Enhancements for ITK-SNAP 3D Medical Image Segmentation Software.

Comprehensive Neuroscience Center Pilot Grant (PI)
01/01/07-12/31/08
University of Pennsylvania Comprehensive Neuroscience Center
Detailed Analysis Of Hippocampal Recruitment For Implicit And Explicit Learning In Aging And Dementia

bibliography

Journal Papers

Full-Length Peer-Reviewed Conference Papers