1 paper · 1 filter
Ishan Rajendrakumar Dave, Simon Jenni, Mubarak Shah
Self-supervised approaches for video have shown impressive results in video understanding tasks. However, unlike early works that leverage temporal self-supervision, current state-…