GeoNet: Unsupervised Learning of Dense Depth, Optical Flow and Camera Pose.
Zhichao YinJianping ShiPublished in: CoRR (2018)
Keyphrases
- camera pose
- unsupervised learning
- optical flow
- camera motion
- visual odometry
- structure from motion
- bundle adjustment
- relative pose
- image sequences
- camera pose estimation
- pose estimation
- point correspondences
- motion parameters
- feature tracking
- feature points
- supervised learning
- computer vision
- single view
- position and orientation
- least squares
- motion estimation
- epipolar geometry
- scene structure
- flow field
- camera calibration
- motion model
- ego motion
- optical flow estimation
- d scene
- dimensionality reduction
- focal length
- semi supervised
- motion segmentation
- moving objects
- motion field
- object recognition
- urban environments
- closed form solutions
- camera parameters
- video sequences
- model selection
- line segments
- multi view
- point cloud
- special case
- d objects
- multiple views