Speaker: Michael Farrell Jr.

Title: An Investigation of Automated Image Registration

Date: October 24, 2003

Time: 3:00 pm

Location: GCATT Room 325

Abstract:
Registering two different images whose subject is the same -- be they from different viewing perspectives, different sensors, or both-- is a very desirable capability in many image processing/computer vision applications. In medicine, doctors can make better diagnoses with registered images of the same organ taken by different sensors such as MRI, CT (x-ray) and PET. Similarly, there are numerous environmental and military applications for remotely sensed imagery that has been registered. This talk will present the issues of image registration, focusing on the challenges of automated multi-sensor image registration. An example of a "real world" automated image registration scheme will be given.

Biography:

Michael Farrell received two undergraduate degrees from the College of Engineering at The University of Iowa and earned his Masters in ECE at Georgia Tech in 2002. Before graduate school he worked in industry for Intel Corporation and Inventa Technologies. At Iowa, he was a Richard Miller Engineering Scholar and won the Clifford Baumback Award for Technologial Entrepreneurship. His current research is in remote sensing, specifically in detection and classification algorithms for hyperspectral imagery. Michael spent the past summer at MIT Lincoln Laboratory working on surface surveillance applications with SAR, IR,
and E-O imagery.

Slides:

Speaker:	Michael Farrell Jr.
Title:	An Investigation of Automated Image Registration
Date:	October 24, 2003
Time:	3:00 pm
Location:	GCATT Room 325
Abstract:	Registering two different images whose subject is the same -- be they from different viewing perspectives, different sensors, or both-- is a very desirable capability in many image processing/computer vision applications. In medicine, doctors can make better diagnoses with registered images of the same organ taken by different sensors such as MRI, CT (x-ray) and PET. Similarly, there are numerous environmental and military applications for remotely sensed imagery that has been registered. This talk will present the issues of image registration, focusing on the challenges of automated multi-sensor image registration. An example of a "real world" automated image registration scheme will be given.
Biography:	Michael Farrell received two undergraduate degrees from the College of Engineering at The University of Iowa and earned his Masters in ECE at Georgia Tech in 2002. Before graduate school he worked in industry for Intel Corporation and Inventa Technologies. At Iowa, he was a Richard Miller Engineering Scholar and won the Clifford Baumback Award for Technologial Entrepreneurship. His current research is in remote sensing, specifically in detection and classification algorithms for hyperspectral imagery. Michael spent the past summer at MIT Lincoln Laboratory working on surface surveillance applications with SAR, IR, and E-O imagery.
Slides: