Research map / 3D vision / Depth and stereo
DENSE RGB SLAM WITH NEURAL IMPLICIT MAPS
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 six feature-volume hierarchies instead of fewer levels (ablation) | Dense RGB SLAM · Replica office-0 | ATE RMSE(cm): 0.67 vs 0.76; 2.27; X | Helped |
| Use six feature-volume hierarchies instead of fewer levels (ablation) | Dense RGB SLAM · Replica office-0 | Accuracy(cm): 2.60 vs 3.39; 6.78 | Helped |
| Increase feature channels from 1 to 2, 4, 8, or 16 while keeping six levels (ablation) | Dense RGB SLAM · Replica office-0 | ATE RMSE(cm): 0.67; 0.63; 0.69; 0.61; 0.66 | No clear effect |
| Increase feature channels from 1 to 4 or 16 at initialization (ablation) | Dense RGB SLAM · Replica office-0 | Depth Error(cm): 1.85; 1.72; 1.93 | No clear effect |
| Remove the MLP decoder and sum level occupancies directly (ablation) | Dense RGB SLAM · Replica office-0 | Accuracy(cm): 2.60 -> 11.43 | Hurt |
| Use multiscale patch warping with patch sizes {1,5,11} instead of {1,11} (ablation) | Dense RGB SLAM initialization · Replica office-0 | Depth Error(cm): 3.73 -> 1.85 | Helped |
| Use pixel-wise warping {1} instead of multiscale patch warping {1,5,11} (ablation) | Dense RGB SLAM initialization · Replica office-0 | Depth Error(cm): 1.85 -> 35 | Hurt |
| Use two-thread rather than one-thread implementation (ablation) | Dense RGB SLAM · Replica | ATE RMSE[cm]: 0.46 -> 0.89 | Hurt |
| Use two-thread rather than one-thread implementation (ablation) | Dense RGB SLAM · Replica | Accuracy[cm]: 4.03 -> 4.29 | Mixed |
Closest papers by meaning
- Benchmarking Implicit Neural Representation and Geometric Rendering in Real-Time RGB-D SLAM (CVPR 2024 · similarity 0.91)
- Loopy-SLAM: Dense Neural SLAM with Loop Closures (CVPR 2024 · similarity 0.89)
- Implicit Event-RGBD Neural SLAM (CVPR 2024 · similarity 0.89)
- IBD-SLAM: Learning Image-Based Depth Fusion for Generalizable SLAM (CVPR 2024 · similarity 0.89)
- Photo-SLAM: Real-time Simultaneous Localization and Photorealistic Mapping for Monocular Stereo and RGB-D Cameras (CVPR 2024 · similarity 0.89)
- Query Quantized Neural SLAM (AAAI 2025 · similarity 0.88)
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.