Broad Area Colloquium For AI-Geometry-Graphics-Robotics-Vision


Image Guided Surgery

W. Eric L. Grimson
AI Laboratory
Massachusetts Institute of Technology

Monday, Apr 23, 2001, 4:15PM
TCSEQ 201
http://robotics.stanford.edu/ba-colloquium/

Abstract

Minimally invasive procedures promise to revolutionize surgery, by allowing surgeons to operate through narrow openings in the body. While these procedures in principle reduce the impact on the patient, they impose a burden on the surgeon, who must operate while only observing through narrow fields of view. To enable surgeons to visualize the entire field, including information not visible to the human eye, we provide a set of computer vision tools. Segmentation methods build patient specific models of anatomy from medical scans, utilizing knowledge of standard shapes of structure and their modes of variance, as well as knowledge of spatial layout of anatomy. Registration methods automatically fuse multi-modal information to a single model, and align such models with actual patient position in the OR. Visualization methods provide augmented reality viewing of the patient, exposing hidden structures to the surgeon, as well as the locations of her instruments relative to critical structures. We will describe our experience in building and deploying such image guided surgical systems, and describe their impact on several hundred actual surgical procedures.

About the Speaker

Eric Grimson received a B.Sc. in Mathematics and Physics from the University of Regina, and a Ph.D. in Mathematics from the Massachusetts Institute of Technology. He is currently Professor of Computer Science and Engineering and Associate Director of the Artificial Intelligence Laboratory at MIT, a Lecturer on Radiology at Harvard Medical School, and holds the Bernard Gordon Chair of Medical Engineering at MIT.

His current research interests include: medical image computation, image guided surgery, activity detection and classification in video, trainable vision systems, image database indexing, and object recognition.


Contact: bac-coordinators@cs.stanford.edu

Back to the Colloquium Page