Frame-wise Action Representations for Long Videos via Sequence Contrastive Learning.

Minghao Chen Fangyun Wei Chong Li Deng Cai

Published in: CoRR (2022)

Keyphrases

learning systems
video frames
learning tasks
human actions
video sequences
prior knowledge
active learning
action sequences
neural network
reinforcement learning
human activities
spatio temporal
mobile robot
space time
motion vectors
inductive inference