Compare with
SpatialTracker
,
SceneTracker
, and the 3D version of
DOT
on 3D dense tracking on in-the-wild videos.
The inference time (in minutes) are denoted on top of each video.