Lumiere: A Space-Time Diffusion Model for Video Generation.
Omer Bar-TalHila CheferOmer TovCharles HerrmannRoni PaissShiran ZadaAriel EphratJunhwa HurYuanzhen LiTomer MichaeliOliver WangDeqing SunTali DekelInbar MosseriPublished in: CoRR (2024)
Keyphrases
- space time
- diffusion model
- video sequences
- spatial and temporal
- input video
- spatio temporal
- video representation
- dynamic scenes
- information diffusion
- human actions
- anisotropic diffusion
- diffusion process
- video data
- scan statistic
- temporal domain
- diffusion tensor
- frame rate
- dynamic textures
- social networks
- key frames
- moving objects
- image processing
- computer vision