An Overview of Medical Imaging Algorithms

Date: 
Wednesday, November 1, 2006 - 17:30
Location: 
TH 331
Presenter: 
Romain Moreau-Gobard, Project Manager at Siemens
Abstract: 
What is the difference between the controversial video game "Grand Theft Auto" and the latest million-dollar software installed in high-tech hospitals around the world, used for diagnosis of life-threatening diseases? 3D Visualization of volumetric medical image data is crucial for diagnosis and surgery pre-planning. Automated tasks performed by computer-aided diagnosis (CAD) algorithms and complex geometric image reformatting are part of every-day tools for radiologists. But behind the sophisticated facade of medical imaging science, some fundamental techniques actually utilize tools from a variety of fun computer science applications. During this talk, I will present different medical software and algorithms used for screening heart conditions, blood vessel diseases, and detecting internal bleeding. We will then correlate those medical imaging techniques with totally different industries, such as 3D modeling, architectural design, and video games.
Bio: 

Romain Moreau-Gobard is a Project Manager at Siemens Corporate Research (Princeton, NJ) in the Imaging and Visualization department. He holds degrees in Artificial Intelligence from EPITA, Paris, and in Business and Project Management from Stevens Institute of Technology, Hoboken, NJ.

Romain's work at Siemens focused on segmenting blood vessels in 3D medical datasets, acquired by computed tomography angiography (CT) and magnetic resonance (MRI) scanners. His research work on heart coronary artery segmentation led to over 15 US and international patents and is integrated into different medical product, helping physicians to perform more accurate diagnosis. He also spent 2 years in Bangalore, India, setting up a research group now composed of 80 engineers and research scientists.

Romain is now working in collaboration with Siemens Medical in Mountain View, on a DARPA grant, developing a portable, light-weight, non-invasive, automated system for the detection, localization, and coagulation of internal bleeding that is operable by minimally trained personnel in the combat environment.