Research map / Continual learning / Online and replay-based
Learn to Memorize and to Forget: A Continual Learning Perspective of Dynamic SLAM
Baicheng Li, Zike Yan, Dong Wu, Hanqing Jiang, Hongbin Zha
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 |
|---|---|---|---|
| Mask dynamic instances with a continually learned classifier before tracking and mapping | Co-SLAM · Bonn crowd | ATE RMS (m): fail -> 0.116 | Improved |
| Mask dynamic instances with the classifier | Co-SLAM · Bonn synchronous | ATE RMS (m): 0.634 -> 0.130 | Improved |
| Compare against DynaSLAM on mild-motion sequences | Ours · Bonn balloon and kidnapping_box | ATE RMS (m): 0.050 (DynaSLAM) vs 0.206; 0.026 vs 0.112 | Hurt |
| Add replay buffer for classifier training (ablation) | Motion classifier · Bonn person_tracking and balloon | classifier update count | Improved |
| Add classifier replay to avoid forgetting dynamic instances (ablation) | Motion classifier · TUM fr3_walking_static frame 190 | motion segmentation (qualitative) | Improved |
| Use BLIP-2 rather than DINOv2 as visual encoder (ablation) | Motion classifier · Bonn balloon | ATE RMS (m): 0.228 -> 0.206 | Improved |
| Use CLIP instead of BLIP-2 (ablation) | Motion classifier · Bonn person_tracking | ATE RMS (m): 0.274 -> 0.271 | No clear effect |
| Add pre-trained human prior to classifier (ablation) | Motion classifier · Bonn person_tracking | ATE RMS (m): 0.274 -> 0.259 | Improved |
| Add pre-trained human prior to classifier (ablation) | Motion classifier · Bonn balloon | ATE RMS (m): 0.206 -> 0.212 | Hurt |
| Add position feature to classifier input (ablation) | Motion classifier · kidnapped box scene | map update (qualitative) | Improved |
Closest papers by meaning
- Continual Learning for Image-Based Camera Localization (ICCV 2021 · similarity 0.86)
- Change-Robust Online Topological Memory for Long-Term Relocalization and Semantic Navigation (NeurIPS 2026 · similarity 0.86)
- Continual Neural Mapping: Learning an Implicit Scene Representation From Sequential Observations (ICCV 2021 · similarity 0.86)
- MemoNav: Working Memory Model for Visual Navigation (CVPR 2024 · similarity 0.85)
- C-NAV: Towards Self-Evolving Continual Object Navigation in Open World (NeurIPS 2025 · similarity 0.84)
- Query Quantized Neural SLAM (AAAI 2025 · 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.
