Net: Semantic-Aware Self-supervised Depth Estimation with Monocular Videos and Synthetic Data.
Bin ChengInderjot Singh SagguRaunak ShahGaurav BansalDinesh BharadiaPublished in: ECCV (30) (2020)
Keyphrases
- synthetic data
- depth estimation
- monocular images
- dynamic scenes
- depth cues
- stereo vision
- depth map
- depth perception
- depth estimates
- stereo matching
- scene understanding
- depth information
- moving camera
- image sequences
- video sequences
- multi view
- real scenes
- stereo pair
- real image data
- data sets
- space time
- video frames
- real world
- video surveillance
- super resolution
- object recognition
- moving objects
- feature matching
- background subtraction
- pose estimation
- object detection
- stereo images
- video data
- matching algorithm
- dynamic programming
- motion segmentation
- high quality
- d scene
- real time