Unsupervised Monocular Depth Estimation CNN Robust to Training Data Diversity.
Valery AnisimovskiyAndrey ShcherbininSergey TurkoIlya KurilinPublished in: Canadian Conference on AI (2020)
Keyphrases
- depth estimation
- monocular images
- training data
- depth map
- depth cues
- stereo vision
- real scenes
- depth information
- stereo matching
- data sets
- image sequences
- optical flow
- learning algorithm
- pose estimation
- vision system
- model selection
- super resolution
- point cloud
- partial occlusion
- scene understanding
- semi supervised
- pairwise
- training set
- object recognition
- depth estimates