Research map / Learning theory and algorithms / Prediction and decision losses
Spline Error Weighting for Robust Visual-Inertial Fusion
Hannes Ovrén, Per-Erik Forssén
Read the paperCheck 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 |
|---|---|---|---|
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Bodycam 1 | EPE [m]: 37.95 -> 0.22 | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Bodycam 1 | scale error [%]: 124.2% -> 3.4% | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Bodycam 2 | EPE [m]: 27.86 -> 0.40 | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Bodycam 2 | scale error [%]: 423.1% -> 8.7% | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Handheld 1 | EPE [m]: 0.52 -> 0.30 | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Handheld 1 | scale error [%]: 19.0% -> 1.4% | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Handheld 2 | EPE [m]: 0.94 -> 0.56 | Improved |
| Replace inverse noise-covariance weighting and fixed knot spacing with SEW-selected weights and knot spacing | Continuous-time visual-inertial SfM · Handheld 2 | scale error [%]: 1.7% -> 2.0% | Hurt |
| Scale the SEW-derived IMU residual weights (ablation) | Continuous-time visual-inertial SfM · Real sequences | endpoint error and scale error | Mixed |
| Choose knot spacing using requested signal-retained quality | IMU trajectory splines · Bodycam and Handheld datasets | obtained spline quality | Improved |
Closest papers by meaning
- SplineGS: Robust Motion-Adaptive Spline for Real-Time Dynamic 3D Gaussians from Monocular Video (CVPR 2025 · similarity 0.85)
- A Robust Method for Strong Rolling Shutter Effects Correction Using Lines With Automatic Feature Selection (CVPR 2018 · similarity 0.84)
- Gaussian Splatting on the Move: Blur and Rolling Shutter Compensation for Natural Camera Motion (ECCV 2024 · similarity 0.84)
- SplArt: Articulation Estimation and Part-Level Reconstruction with 3D Gaussian Splatting (ICCV 2025 · similarity 0.83)
- An N-Point Linear Solver for Line and Motion Estimation with Event Cameras (CVPR 2024 · similarity 0.83)
- Scalable Benchmarking and Robust Learning for Noise-Free Ego-Motion and 3D Reconstruction from Noisy Video (ICLR 2025 · similarity 0.83)
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.
