BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Motion Tracking - Fabian Wauthier\, University of Edinburgh
DTSTART:20070316T100000Z
DTEND:20070316T110000Z
UID:TALK6836@talks.cam.ac.uk
CONTACT:Oliver Williams
DESCRIPTION:The subject of the presentation is concerned with online track
 ing of moving foreground objects through an image sequence. As part of the
  fourth year honours project\, relevant parts of the paper ``Learning patt
 erns of activity using real-time tracking`` (Stauffer & Grimson\, 2000) we
 re implemented and evaluated. Specifically\, an adaptive background subtra
 ction technique\, based on pixel-wise Mixtures of Gaussians is used to est
 imate a multimodal distribution over background pixels. The mixture repres
 entation allows repetitive background motion to be absorbed into the backg
 round estimate. Foreground objects in the scene become apparent as ``blobs
 `` using outlier detection on each pixel`s mixture. To track these blobs a
 cross several image frames\, Kalman filtering is employed. In the context 
 of tracking multiple objects simultaneously\, determining the association 
 between detected blobs and tracked objects is a non-trivial problem\, not 
 fully described by Stauffer and Grimson. The association problem is formul
 ated as an instance of the linear assignment problem and solved using exis
 ting algorithms. The results indicate that multiple objects can be tracked
  in simple scenes reasonably well. 
LOCATION:Small public lecture room\, Microsoft Research Ltd\, 7 J J Thomso
 n Avenue (Off Madingley Road)\, Cambridge
END:VEVENT
END:VCALENDAR