Gary Christensen- Colloquium Speaker

Professor Electrical and Computer Engineering & Radiation Oncology
Date: 
Thursday, September 25, 2014 - 3:00pm
Colloquium Title: 
Modeling Mechanical Properties of the Lung using Image Registration and the Shape Collapse Problem in Image Registration
Location: 
Reception at 3:00 p.m. in 241 SH / Talk at 3:30 in 61 SH

This talk will consist of two parts. The first part will explore using non-rigid image registration to characterize the mechanical properties of the lung and the second part will explore the shape collapse problem in non-rigid image registration.

Image registration plays an important role in pulmonary image analysis. Accurate image registration is a challenging problem when the lungs have undergone large distance deformation. Registration results estimate the local tissue movement and are useful for studying lung mechanical properties. This talk will discuss non-rigid registration of CT image volumes of the lung using the sum of square tissue volume differences similarity cost and using a feature-based vesselness constraint.  Results will be presented characterizing the accuracy of the method.  Next, we will discuss how the estimated registrations can be used to characterize the mechanical properties of the lung.  Finally, we will discuss the challenging registration problem of aligning intra-subject CT acquired during radiotherapy in which "infiltrative" tumors regress over the course of treatment.

The second part of the talk will discuss the shape collapse problem in non-rigid image registration. The shape collapse problem is the situation when an appendage of a deforming object does not overlap with the target shape and collapses to a set of zero measure during the registration process. The dual problem occurs when a new appendage grows out of the object to match the target shape. In both cases, the estimated correspondence between the source and target objects is often undesirable.  The shape collapse problem is caused by deforming the moving image in the gradient direction of the similarity cost and affects both small and large deformation registration algorithms. Minimizing a registration cost function by following the similarity-cost gradient drives the registration to a local energy minimum and does not permit an increase in energy to ultimately reach a lower energy state. Furthermore, once an object collapses locally, it has zero measure under the similarity cost in this region and is permanently stuck in a local minimum. This talk will present a criterion for detecting image regions that will collapse if the similarity cost gradient direction is followed during optimization. This criterion is based on the skeletal points of the moving image in the symmetric difference of the original two binary images. Experimental results are presented that demonstrate that the shape collapse problem can be detected before registration.

 

bio:

Gary E. Christensen received his BS degrees in electrical engineering and computer science, graduating magna cum laude, in 1988 and his MS and DSc degrees in electrical engineering in 1989 and 1994, respectively, from Washington University, St. Louis. In 1997, Dr. Christensen joined the faculty of the Department of Electrical and Computer Engineering, the University of Iowa, where he is currently holds the rank of professor.  He has a joint appointment in the Department of Radiation Oncology and is a member of the Iowa Institute for Biomedical Imaging, Iowa Comprehensive Lung Imaging Center, and Holden Comprehensive Cancer Center.  Previously, Dr. Christensen was an assistant professor with the Washington University School of Medicine from 1994 to 1996 with joint appointments in the Department of Surgery, Mallinckrodt Institute of Radiology, and Department of Electrical Engineering and directed the Craniofacial Imaging Laboratory, St. Louis Children’s Hospital, Washington University Medical Center. He has published over 100 scientific papers and is the co-inventor on four patents. His primary research interests include image registration, 3-D visualization, medical imaging, and deformable shape models.