Anthony Hoogs

Director of Computer Vision

Kitware, Inc.

   "Getting Results: Video Analysis with a Purpose"

In the 1980’s, Ruzena, Aloimonos and others introduced the idea of coupling computer vision with the underlying task, purpose or intent of an active agent, and called this “active vision.” The related concept of “functional” computer vision was developed by Stark and Bowyer a few years later, in an attempt to recognize objects by their function or purpose inferred from single images. Recently, we have been revisiting the concepts of function, intent and purpose in computer vision through video analysis. Video enables the direct observation of object behaviors and interactions over time, from which purpose, function and intent can be recognized much more easily than in single, static images. I will present a summary of our investigations in several related areas, including the learning and recognition of complex activities involving multiple people, demonstrated on American football plays; unsupervised learning of functional scene element categories in webcam video; and learning normal behavior patterns in order to detect anomalies in surveillance video.

Wednesday, April 22nd, 2009
8:30 am - 5:30 pm

School of Engineering and Applied Science

University of Pennsylvania

210 South 33rd Street

Berger Auditorium

Skirkanich Hall

GRASP Lab - Levine Hall 4th Floor

For more information regarding our speaker please visit:

http://www.cc.gatech.edu/~hic/8803-Mobile-08/Welcome.html