4 ms·
From this paper http://info.ee.surrey.ac.uk/Personal/Z.Kalal/Publications/2010_icip.pdf http://info.ee.surrey.ac.uk/Personal/Z.Kalal/Publications/20... it seems
by karl_nerd 16y ago
From this paper http://info.ee.surrey.ac.uk/Personal/Z.Kalal/Publications/2010_icip.pdf http://info.ee.surrey.ac.uk/Personal/Z.Kalal/Publications/20... it seems that the system collects images of successful detections, and that way gets more data to base its' coming detections on. In this demo this seem more responsive than Kinect, but on the other hand he's only tracking one object at the time, is he?
- regularfry 16y agoI think the single-object constraint here is an artifact of the UI rather than a limitation of the algorithm.