Computer Vision Topics for Interaction

Speaker
Martial Hebert
Professor, Robotics Institute, Carnegie Mellon University

When
-

Where
Newell-Simon Hall 1305 (Michael Mauldin Auditorium)

Description

In this talk, I will review a few topics in computer vision that are directly relevant to various applications involving interaction with the environment or with people. Techniques for object recognition are used for locating objects in a scene observed by a camera. Techniques for place recognition are used for relating the scene currently imaged (for example, John’s office) with a large collection of images acquired over a period of time.Techniques for event detection can be used to identify atomic actions from agents observed in a video camera (for example, standing up, waving, picking something up, etc.). Finally, techniques for self activity recognition are used to estimate the type of activity in which the subject carrying the vision system is engaged. The basic approaches as well as examples of what can (and what cannot!) be done will be shown for each technique. Finally, the relations of the basic computer vision with various application such as intelligent mobility, virtual coaches, and safe driving will be discussed.

Speaker's Bio

Martial Hebert is a professor at the Robotics Institute, Carnegie Mellon University. His interests include computer vision, especially object recognition, scene understanding, and video analysis, processing of 3-D data for building 3-D models, recognizing objects, and for autonomous navigation, and perception for autonomous mobile systems. His current projects include the development of techniques for recognizing image categories in images, for detecting common events in video sequences, and for building 3-D representation of dynamic environments for unmanned mobile systems.