BuilDiff: 3D Building Shape Generation using Single-Image Conditional Point Cloud Diffusion Models.
Yao WeiGeorge VosselmanMichael Ying YangPublished in: CoRR (2023)
Keyphrases
- single image
- point cloud
- diffusion models
- shape recovery
- structure from motion
- multiple images
- object surface
- diffusion model
- surface reconstruction
- intrinsic images
- d scene
- image pairs
- information diffusion
- light source
- point sets
- shape from shading
- social networks
- super resolution
- social media
- surface orientation
- machine learning
- image sequences
- vanishing points
- shape prior
- vector field
- three dimensional
- image data
- optical flow