Research map / 3D vision / Depth and stereo
Pseudo RGB-D for Self-Improving Monocular SLAM and Depth Prediction
Lokender Tiwari, Pan Ji, Quoc-Huy Tran, Bingbing Zhuang, Saket Anand, Manmohan Chandraker
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 |
|---|---|---|---|
| add depth consistency loss from SLAM depths (ablation) | pRGBD-Refined · KITTI Eigen (ablation, 1st self-improving loop) | RMSE: 4.956 -> 4.809 | Improved |
| add symmetric depth transfer loss (ablation) | pRGBD-Refined · KITTI Eigen (ablation, 1st self-improving loop) | RMSE: 4.867 -> 4.809 | Improved |
| add photometric loss (ablation) | pRGBD-Refined · KITTI Eigen (ablation, 1st self-improving loop) | RMSE: 4.855 -> 4.809 | Improved |
| add depth consistency loss (ablation) | pRGBD-Refined · KITTI Eigen (ablation, 1st self-improving loop) | Abs Rel: 0.117 -> 0.117 | No clear effect |
| add depth consistency loss (ablation) | pRGBD-Refined · KITTI Eigen (ablation, 1st self-improving loop) | a1: 0.862 -> 0.873 | Improved |
| replace monocular depth with pseudo RGB-D input to ORB-SLAM (pRGBD-Initial vs RGB ORB-SLAM) | ORB-SLAM · KITTI Odometry seq 09 | RMSE (ATE): 18.34 -> 12.21 | Improved |
| add self-improving depth refinement then re-run SLAM (Refined vs Initial) | pRGBD-SLAM · KITTI Odometry seq 10 | RMSE (ATE): 8.30 -> 6.35 | Improved |
| pseudo RGB-D instead of RGB ORB-SLAM | ORB-SLAM · KITTI Odometry seq 09 | Rel Rot: 0.004 -> 0.011 | Hurt |
| add depth refinement via self-improving loops to MonoDepth2-M | MonoDepth2-M · KITTI Eigen split test | Sq Rel: 0.941 -> 0.793 | Improved |
| add depth refinement on TUM freiburg3 | pRGBD-Initial/MonoDepth2-M · TUM freiburg3 depth | RMSE: 1.090 -> 0.743 | Improved |
| add pose refinement on TUM walking_xyz | pRGBD-Initial · TUM freiburg3 walking_xyz pose | RMSE: 0.23 -> 0.09 | Improved |
| pRGBD-Refined vs stereo-trained Monodepth2-S | MonoDepth2 · KITTI Eigen split test | Abs Rel: 0.109 -> 0.113 | Mixed |
Closest papers by meaning
- Depth2Pose: A Pose-Based Benchmark for Monocular Depth Estimation without Ground-Truth Depth (NeurIPS 2026 · similarity 0.89)
- IBD-SLAM: Learning Image-Based Depth Fusion for Generalizable SLAM (CVPR 2024 · similarity 0.88)
- Photo-SLAM: Real-time Simultaneous Localization and Photorealistic Mapping for Monocular Stereo and RGB-D Cameras (CVPR 2024 · similarity 0.88)
- SLAM3R: Real-Time Dense Scene Reconstruction from Monocular RGB Videos (CVPR 2025 · similarity 0.88)
- Loopy-SLAM: Dense Neural SLAM with Loop Closures (CVPR 2024 · similarity 0.87)
- VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifold (NeurIPS 2025 · similarity 0.87)
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.
