Back to the Feature: Learning Robust Camera Localization from Pixels to Pose
Paper in proceeding, 2021
Camera pose estimation in known scenes is a 3D geometry task recently tackled by multiple learning algorithms. Many regress precise geometric quantities, like poses or 3D points, from an input image. This either fails to generalize to new viewpoints or ties the model parameters to a specific scene. In this paper, we go Back to the Feature: we argue that deep networks should focus on learning robust and invariant visual features, while the geometric estimation should be left to principled algorithms. We introduce PixLoc, a scene-agnostic neural network that estimates an accurate 6-DoF pose from an image and a 3D model. Our approach is based on the direct alignment of multiscale deep features, casting camera localization as metric learning. PixLoc learns strong data priors by end-to-end training from pixels to pose and exhibits exceptional generalization to new scenes by separating model parameters and scene geometry. The system can localize in large environments given coarse pose priors but also improve the accuracy of sparse feature matching by jointly refining keypoints and poses with little overhead. The code will be publicly available at github.com/cvg/pixloc.
Author
Paul-Edouard Sarlin
Swiss Federal Institute of Technology in Zürich (ETH)
Ajaykumar Unagar
Swiss Federal Institute of Technology in Zürich (ETH)
Mans Larsson
Eigenvision AB
Hugo Germain
École des Ponts ParisTech
Carl Toft
Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering
Viktor Larsson
Swiss Federal Institute of Technology in Zürich (ETH)
Marc Pollefeys
Microsoft Corporation
Swiss Federal Institute of Technology in Zürich (ETH)
Vincent Lepetit
École des Ponts ParisTech
Lars Hammarstrand
Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering
Fredrik Kahl
Computer vision and medical image analysis
Torsten Sattler
Chalmers, Electrical Engineering, Signal Processing and Biomedical Engineering
Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
10636919 (ISSN)
3246-3256978-1-6654-4509-2 (ISBN)
Online, ,
Subject Categories
Bioinformatics (Computational Biology)
Media Engineering
Computer Vision and Robotics (Autonomous Systems)
DOI
10.1109/CVPR46437.2021.00326