Research map / 3D vision / 3D generation
NICE-SLAM: Neural Implicit Scalable Encoding for SLAM
Check 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 hierarchical feature grids with pretrained decoders instead of a single scene MLP | NICE-SLAM · Replica | Depth L1: 7.64 -> 3.53 | Helped |
| Use hierarchical feature grids with pretrained decoders instead of a single scene MLP | NICE-SLAM · ScanNet | ATE RMSE: 36.67 -> 9.63 | Helped |
| Use hierarchical feature grids with shallow decoders | NICE-SLAM · Point query | FLOPs [×10^3]: 443.91 -> 104.16 | Helped |
| Use hierarchical feature grids with shallow decoders | NICE-SLAM · Tracking | runtime [ms]: 101 -> 47 | Helped |
| Add Local BA (ablation) | NICE-SLAM · ScanNet | ATE RMSE: 37.74 -> 9.63 | Helped |
| Add the photometric loss and color representation (ablation) | NICE-SLAM · ScanNet | ATE RMSE: 32.02 -> 9.63 | Helped |
| Select keyframes with visual overlap instead of using iMAP keyframes (ablation) | NICE-SLAM · ScanNet | ATE RMSE: 12.10 -> 9.63 | Helped |
| Use 3 feature-grid levels instead of 2 (ablation) | NICE-SLAM · Replica room-0 | Acc. [cm]: 2.87 -> 2.78 | Helped |
| Use 4 feature-grid levels instead of 3 (ablation) | NICE-SLAM · Replica room-0 | Acc. [cm]: 2.78 -> 3.15 | Hurt |
| Increase mapping iterations from 15 to 60 (ablation) | NICE-SLAM · Replica room-0 | Depth L1 [cm]: 2.31 -> 1.87 | Helped |
| Increase mapping iterations from 60 to 240 (ablation) | NICE-SLAM · Replica room-0 | Depth L1 [cm]: 1.87 -> 1.59 | Helped |
| Filter pixels with large depth/color re-rendering loss | NICE-SLAM · Co-Fusion | ATE RMSE [cm]: 7.8 -> 1.6 | Helped |
Closest papers by meaning
- SNI-SLAM: Semantic Neural Implicit SLAM (CVPR 2024 · similarity 0.89)
- Query Quantized Neural SLAM (AAAI 2025 · similarity 0.89)
- Benchmarking Implicit Neural Representation and Geometric Rendering in Real-Time RGB-D SLAM (CVPR 2024 · similarity 0.88)
- IBD-SLAM: Learning Image-Based Depth Fusion for Generalizable SLAM (CVPR 2024 · similarity 0.88)
- Loopy-SLAM: Dense Neural SLAM with Loop Closures (CVPR 2024 · similarity 0.87)
- DeepPointMap: Advancing LiDAR SLAM with Unified Neural Descriptors (AAAI 2024 · 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.