Research map / Architectures and efficiency / RNNs and normalization
Information-Driven Direct RGB-D Odometry
Alejandro Fontan, Javier Civera, Rudolph Triebel
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 |
|---|---|---|---|
| Select points by pose-information gain, with grid pre-selection (ablation) | ID-RGBDO point selection · TUM RGB-D sequences | K2FE | Improved |
| Select points by pose-information gain instead of random selection (ablation) | ID-RGBDO point selection · TUM RGB-D sequences | K2FE | Improved |
| Select points by pose-information gain instead of grid selection (ablation) | ID-RGBDO point selection · TUM RGB-D sequences | K2FE | Improved |
| Use information-based selection with grid pre-selection rather than information-only selection (ablation) | ID-RGBDO point selection · TUM RGB-D sequences | K2FE | No clear effect |
| Reduce tracking to 24 informative points per keyframe (ablation) | ID-RGBDO · TUM RGB-D sequences | computational demand: 10× | Improved |
| Reduce the number of tracked points from hundreds to 24 informative points (ablation) | ID-RGBDO · TUM RGB-D sequences | trajectory accuracy | No clear effect |
| Set keyframe creation using relative pose-entropy reduction (ablation) | ID-RGBDO · TUM RGB-D sequences | RPE and ATE | Improved |
| Use information-based keyframe marginalization to remove redundant keyframes (ablation) | ID-RGBDO · TUM RGB-D sequences | trajectory accuracy | No clear effect |
| Use 24 informative points and 8 keyframes in sliding-window bundle adjustment (ablation) | ID-RGBDO · TUM RGB-D sequences | bundle-adjustment cost: 10× | Improved |
| Use ID-RGBDO instead of the strongest reported baseline | RGB-D odometry · TUM RGB-D sequences | RPE and ATE | Mixed |
Closest papers by meaning
- VOLDOR: Visual Odometry From Log-Logistic Dense Optical Flow Residuals (CVPR 2020 · similarity 0.87)
- Direct Sparse Odometry (TPAMI 2017 · similarity 0.87)
- RGB-D SLAM in Dynamic Environments Using Point Correlations (TPAMI 2020 · similarity 0.86)
- Beyond Tracking: Selecting Memory and Refining Poses for Deep Visual Odometry (CVPR 2019 · similarity 0.86)
- Information-Theoretic Odometry Learning (IJCV 2022 · similarity 0.86)
- LEAP-VO: Long-term Effective Any Point Tracking for Visual Odometry (CVPR 2024 · 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.
