Multi-Frame Self-Supervised Depth Estimation with Multi-Scale Feature Fusion in Dynamic Scenes.
Jiquan ZhongXiaolin HuangXiao YuPublished in: CoRR (2023)
Keyphrases
- depth estimation
- dynamic scenes
- multi frame
- multiscale
- optic flow
- motion estimation
- image sequences
- motion segmentation
- video sequences
- point correspondences
- multi view
- super resolution
- multiple views
- background subtraction
- space time
- motion analysis
- scale space
- moving objects
- optical flow
- video frames
- frame rate
- foreground objects
- image processing
- image segmentation
- flow field
- depth map
- vector field
- camera motion
- image analysis
- object recognition