Tracking Software Available

[hager@CS.YALE.EDU]

We are announcing the first release of our software system for visual
tracking.  The major features of this system are:

   - extremely fast detection and tracking of low-level features 
     such as contrast edges for gray-scale patterns.

   - a simple framework for combining simple features to track complex
     targets.

   - no specialized hardware (other than a framegrabber) required.

The system is written in C++.  The current release (0.95a) contains
trackers for edges, corners, and contours.  It will soon also do SSD
tracking.  It supports the Imaging Technologies FG-101 framegrabber,
and most SGI Indigo video devices, and is fairly easy to port to most
other video devices.

The system has been in use locally for over a year now, and seems to
be fairly stable.  We are interested in getting feedback from other
users, as well as hopefully inspiring more work in real-time vision.

If you want to see a bit more about the system, you can read about it
online at the following URL:

http://www.cs.yale.edu/HTML/YALE/CS/AI/VisionRobotics/YaleTracking.html

This page also contains pointers to a more detailed document about the
system.

The software itself is available by sending mail to hager-greg@cs.yale.edu.

Greg Hager
Yale University

------------------------------