Geometric loss functions for camera pose regression with deep learning.
Alex KendallRoberto CipollaPublished in: CoRR (2017)
Keyphrases
- loss function
- deep learning
- camera pose
- reproducing kernel hilbert space
- support vector
- geometric constraints
- structure from motion
- position and orientation
- unsupervised learning
- pose estimation
- pairwise
- bundle adjustment
- point correspondences
- camera motion
- feature points
- machine learning
- single view
- urban environments
- focal length
- camera calibration
- model selection
- weakly supervised
- mental models
- camera parameters
- closed form solutions
- least squares
- active learning
- vision system
- face recognition