The Computing of X-ray Computed Tomography (CT)

Colloq: Speaker: 
Dr. Thomas Benson
Colloq: Speaker Institution: 
GE Global Research
Colloq: Date and Time: 
Mon, 2010-10-25 10:00
Colloq: Location: 
5100, Room 128 JICS Lecture Hall
Colloq: Host: 
Dr. Jeffrey Vetter
Colloq: Host Email: 
vetter@ornl.gov
Colloq: Abstract: 
Computed tomography (CT) and other medical imaging applications have long been consumers of high-performance computing hardware. Computationally demanding reconstruction algorithms in combination with rapidly increasing data sizes and aggressive performance requirements have contributed to continued interest in the medical imaging community on emerging computing platforms. Furthermore, research continues on iterative reconstruction algorithms where the computational complexity for medical exams is only recently becoming manageable on affordable hardware. In this seminar, I will primarily discuss methods of parallelizing iterative reconstruction algorithms to improve their performance on commodity hardware. I will further discuss research in the application of the Cell Broadband Engine Architecture and GPU platforms for solving computationally demanding CT challenges.
Colloq: Speaker Bio: 
Thomas Benson received a BS in Mathematics and Computer Science in 2001, an MS in Computer Science in 2003, and a PhD in Computer Science in 2006, all from the University of Tennessee, Knoxville. His graduate research focused on the parallelization of iterative reconstruction algorithms for computed tomography reconstruction. Since 2006, he has worked in the CT and X-ray Systems and Applications Laboratory at GE Global Research in Niskayuna, NY. His research focuses on CT algorithm development for medical and security applications as well as performance analysis and improvement for such algorithms on emerging computing platforms.