Efficient Video Diffusion Models via Content-Frame Motion-Latent Decomposition.

Sihyun Yu Weili Nie De-An Huang Boyi Li Jinwoo Shin Anima Anandkumar

Published in: ICLR (2024)

Keyphrases

key frames
successive frames
video frames
input video
space time
diffusion models
temporal continuity
single frame
temporal coherence
multimedia
video data
adjacent frames
motion estimation
video objects
image sequences
video sequences
social networks
moving objects
video content
motion field
image processing