DistInit: Learning Video Representations without a Single Labeled Video.

Rohit Girdhar Du Tran Lorenzo Torresani Deva Ramanan

Published in: CoRR (2019)

Keyphrases

video streams
multimedia
video frames
learning algorithm
video data
video sequences
interactive video
video clips

real time
learning process
space time
learning systems
supervised learning
online video
temporal information

video analysis
reinforcement learning
real time video
video images
digital video
video content
key frames

learning tasks
qualitative models
event detection
multiple representations
knowledge acquisition
computer vision
prior knowledge