Homography-Based Loss Function for Camera Pose Regression.
Clémentin BoittiauxRicard MarxerClaire DuneAurélien ArnaubecVincent HugelPublished in: IEEE Robotics Autom. Lett. (2022)
Keyphrases
- loss function
- camera pose
- feature points
- camera motion
- intrinsic parameters
- reproducing kernel hilbert space
- gradient boosting
- support vector
- point correspondences
- point features
- camera pose estimation
- pose estimation
- pairwise
- structure from motion
- fundamental matrix
- focal length
- bundle adjustment
- epipolar geometry
- image sequences
- position and orientation
- camera parameters
- camera calibration
- optical flow
- urban environments
- model selection
- line segments
- ground plane
- single view
- point sets
- line features
- training data
- d scene
- closed form solutions
- linear regression
- least squares
- scene structure
- projective transformations
- video sequences
- ridge regression
- computer vision
- convex loss functions