Research map / 3D vision / Depth and stereo
The Probabilistic Normal Epipolar Constraint for Frame-to-Frame Rotation Optimization Under Uncertain Feature Positions
Dominik Muhle, Lukas Koestler, Nikolaus Demmel, Florian Bernard, Daniel Cremers
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use PNEC uncertainty-weighted residuals instead of equally weighted NEC residuals | Frame-to-frame relative pose estimation · Synthetic omnidirectional and pinhole camera experiments | rotation error: 0.11 -> 0.08; 0.25 -> 0.20 | Improved |
| Use PNEC uncertainty-weighted residuals instead of equally weighted NEC residuals | Frame-to-frame relative pose estimation · Synthetic omnidirectional camera, noise 0.5 px, with translation | translation error: 1.90 -> 1.29 | Improved |
| Use KLT-PNEC instead of KLT-NEC with the same KLT tracks | Rotation-only visual odometry · KITTI odometry, excluding seq. 01 | RPE_1: 10% | Improved |
| Use KLT-PNEC instead of KLT-NEC with the same KLT tracks | Rotation-only visual odometry · KITTI odometry, excluding seq. 01 | RPE_n: 19% | Improved |
| Use KLT-PNEC instead of KLT-NEC with the same KLT tracks | Rotation-only visual odometry · KITTI odometry, all sequences | RPE_1 and RPE_n | Mixed |
| Use KLT tracks instead of ORB features | NEC-based relative rotation estimation · KITTI odometry | RPE | Improved |
| Use KLT-PNEC instead of KLT-NEC | Rotation-only visual odometry · KITTI odometry | total frame processing time: 53 -> 77 milliseconds | Hurt |
Closest papers by meaning
- Uncertainty-Aware Camera Pose Estimation From Points and Lines (CVPR 2021 · similarity 0.88)
- Relative Pose from SIFT Features (ECCV 2022 · similarity 0.87)
- Learning Correspondence Uncertainty via Differentiable Nonlinear Least Squares (CVPR 2023 · similarity 0.86)
- Globally Optimal Relative Pose Estimation With Gravity Prior (CVPR 2021 · similarity 0.86)
- From Correspondences to Pose: Non-minimal Certifiably Optimal Relative Pose without Disambiguation (CVPR 2024 · similarity 0.86)
- Optimal Relative Pose With Unknown Correspondences (CVPR 2016 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
