CVSformer: Cross-View Synthesis Transformer for Semantic Scene Completion.
Haotian DongEnhui MaLubo WangMiaohui WangWuyuan XieQing GuoPing LiLingyu LiangKairui YangDi LinPublished in: ICCV (2023)
Keyphrases
- view synthesis
- d scene
- uncalibrated images
- image based rendering
- scene geometry
- stereo reconstruction
- reference images
- depth map
- virtual view
- multi view
- depth estimation
- photorealistic
- free viewpoint video
- multiple cameras
- intermediate views
- single image
- depth information
- scene understanding
- three dimensional
- video sequences
- stereoscopic images
- image space
- multiple objects
- optical flow
- real scenes
- ground plane
- multiple images
- dynamic scenes
- stereo matching
- image sequences
- complex scenes
- computer graphics
- vision system