Added: Mar 19, 2010
From: IVMSCIBRVIDEOS
Duration: 0:16
We present a method for unifying tracking and videocontent recognition with applications to Mobile AugmentedReality (MAR). We do so by introducing theRadial Gradient Transform (RGT) and an approximateRGT, yielding the Rotation-Invariant, Fast Feature(RIFF) descriptor. We then demonstrate a method fortracking these descriptors in realtime, to obtain a globalaffine motion model. With this temporal informationthe robustness of video content recognition can be increasedat no additional cost.
Channel: Tech
Rating: ( ratings) Views: 65 Comments: 0
