ConceptExpress: Harnessing Diffusion Models for Single-image Unsupervised Concept Extraction.
Shaozhe HaoKai HanZhengyao LvShihao ZhaoKwan-Yee K. WongPublished in: CoRR (2024)
Keyphrases
- single image
- diffusion models
- shape recovery
- multiple images
- d scene
- diffusion model
- light source
- super resolution
- semi supervised
- indoor scenes
- shape from shading
- partially occluded
- information diffusion
- intrinsic image decomposition
- intrinsic images
- information from multiple images
- three dimensional
- point correspondences
- multiscale