Occlusion-Aware Unsupervised Learning of Monocular Depth, Optical Flow and Camera Pose with Geometric Constraints.
Qianru TengYimin ChenChen HuangPublished in: Future Internet (2018)
Keyphrases
- geometric constraints
- camera pose
- visual odometry
- optical flow
- camera motion
- epipolar geometry
- ego motion
- camera calibration
- image sequences
- planar surfaces
- multiple images
- depth cues
- depth map
- feature correspondences
- motion segmentation
- depth information
- geometric consistency
- structure from motion
- focal length
- computer vision
- motion estimation
- d scene
- motion estimates
- bundle adjustment
- pose estimation
- multi view
- depth images
- point correspondences
- least squares
- ground plane
- single image
- feature points
- camera parameters
- stereo matching
- motion field
- high quality
- machine learning