Boundary-induced and scene-aggregated network for monocular depth prediction.
Feng XueJunfeng CaoYu ZhouFei ShengYankai WangAnlong MingPublished in: CoRR (2021)
Keyphrases
- depth cues
- d scene
- relative depth
- ego motion
- image sequences
- object boundaries
- depth map
- prediction accuracy
- prediction model
- depth estimation
- stereo camera
- real scenes
- peer to peer
- video sequences
- three dimensional
- depth perception
- radial basis function network
- defocused images
- pose estimation
- depth information
- occlusion boundaries
- single image
- input image
- road vehicles
- light source
- stereoscopic images
- depth ordering
- optical flow
- wireless sensor networks
- scene understanding
- ground plane
- stereo vision
- network structure
- multiple views