Research map / Architectures and efficiency / RNNs and normalization
SLAM-Former: Putting SLAM into One Transformer
Yijun Yuan, Zhuoguang Chen, Kenan Li, Weibang Wang, Minghui Qin, Zhijian Fang, Weicheng Zheng, Hang Zhao
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 |
|---|---|---|---|
| use 224px input instead of 518px (ablation) | SLAM-Former · 7-Scenes runtime | FPS: 15.7 -> 25.8 | Improved |
| train on synthetic multi-room HM3D trajectories (data scaling) vs MASt3R-SLAM | SLAM-Former · HM3D | ATE: 0.655 -> 0.188 | Improved |
| compare against MASt3R-SLAM on EuRoC | SLAM-Former · EuRoC | ATE: 0.164 -> 0.150 | Improved |
| add log KV pooling (ablation) | SLAM-Former · Replica | ATE: 0.030 -> 0.034 | No clear effect |
| add sqrt KV pooling (ablation) | SLAM-Former · Replica | Accuracy: 2.09 -> 2.33 | Hurt |
| add log KV pooling (ablation) | SLAM-Former · 7-Scenes | ATE: 0.042 -> 0.041 | No clear effect |
| add log KV pooling (ablation) | SLAM-Former · TUM | ATE: 0.039 -> 0.044 | Hurt |
| increase backend period T from 10 to 50 keyframes (ablation) | SLAM-Former · 7-Scenes | ATE: 0.042 -> 0.046 | Hurt |
| vary backend period T from 5 to 20 (ablation) | SLAM-Former · 7-Scenes | ATE: 0.044 -> 0.042 -> 0.043 | No clear effect |
| lower keyframe translation threshold from 0.2 to 0.05 (ablation) | SLAM-Former · 7-Scenes | Chamfer: 0.032 -> 0.025 | Improved |
| lower keyframe translation threshold from 0.2 to 0.05 (ablation) | SLAM-Former · 7-Scenes | Accuracy: 0.022 -> 0.015 | Improved |
| lower keyframe translation threshold from 0.2 to 0.05 (ablation) | SLAM-Former · 7-Scenes | ATE: 0.043 -> 0.044 | No clear effect |
Closest papers by meaning
- DeepPointMap: Advancing LiDAR SLAM with Unified Neural Descriptors (AAAI 2024 · similarity 0.85)
- Benchmarking Implicit Neural Representation and Geometric Rendering in Real-Time RGB-D SLAM (CVPR 2024 · similarity 0.85)
- MNE-SLAM: Multi-Agent Neural SLAM for Mobile Robots (CVPR 2025 · similarity 0.85)
- MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors (CVPR 2025 · similarity 0.85)
- MAGiC-SLAM: Multi-Agent Gaussian Globally Consistent SLAM (CVPR 2025 · similarity 0.84)
- Loopy-SLAM: Dense Neural SLAM with Loop Closures (CVPR 2024 · similarity 0.84)
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.
