DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models.
Shengqu CaiEric Ryan ChanSongyou PengMohamad ShahbaziAnton ObukhovLuc Van GoolGordon WetzsteinPublished in: ICCV (2023)
Keyphrases
- single view
- diffusion models
- multi view
- vanishing points
- diffusion model
- perspective projection
- information diffusion
- multiple views
- camera pose
- three dimensional
- semi supervised
- surface reconstruction
- social networks
- moving objects
- single image
- image sequences
- viewpoint
- video sequences
- influence maximization
- man made environments
- range data
- multiscale
- dynamic programming