Research map / 3D vision / Depth and stereo
Pose Invariant Topological Memory for Visual Navigation
Asuto Taniguchi, Fumihiro Sasaki, Ryota Yamashina
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 |
|---|---|---|---|
| replace CNN encoder with rotation-invariant SCNN encoder on omnidirectional images (ablation) | POINT · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.908 ± 0.152 -> 0.932 ± 0.163 | Improved |
| use omnidirectional images instead of first-person view images (with CNN encoder) (ablation) | POINT · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.435 ± 0.471 -> 0.908 ± 0.152 | Improved |
| use contrastive learning with cosine similarity (POINT Omni. + CNN) instead of SPTM with omnidirectional images | SPTM (Omni.) · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.040 ± 0.025 -> 0.908 ± 0.152 | Improved |
| use contrastive learning with cosine similarity (POINT Omni. + CNN) instead of HTM with omnidirectional images | HTM (Omni.) · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.083 ± 0.028 -> 0.908 ± 0.152 | Improved |
| add SCNN encoder to SPTM | SPTM · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.020 ± 0.011 -> 0.033 ± 0.016 | No clear effect |
| add SCNN encoder to HTM | HTM · KITTI-360 (real-world omnidirectional) | mean planning success rate: 0.045 ± 0.037 -> 0.103 ± 0.034 | No clear effect |
| add data augmentation to contrastive learning of the edge predictor (ablation) | POINT · Simulated navigation, changed environmental conditions | planning success rate | Improved |
| use SCNN encoder instead of CNN encoder on omnidirectional images when route direction is opposite (ablation) | POINT · Simulated navigation, opposite route direction | planning success rate | Improved |
| use cosine similarity predictor instead of log-bilinear predictor (ablation) | POINT · Simulated navigation, all environmental conditions | planning success rate | Improved |
| use cosine similarity instead of inner product predictor (ablation) | POINT · Simulated navigation, obstacles condition | planning success rate | No clear effect |
Closest papers by meaning
- Hallucinative Topological Memory for Zero-Shot Visual Planning (ICML 2020 · similarity 0.90)
- Change-Robust Online Topological Memory for Long-Term Relocalization and Semantic Navigation (NeurIPS 2026 · similarity 0.89)
- MemoNav: Working Memory Model for Visual Navigation (CVPR 2024 · similarity 0.88)
- Learning with a Mole: Transferable latent spatial representations for navigation without reconstruction (ICLR 2024 · similarity 0.87)
- Kinaema: a recurrent sequence model for memory and pose in motion (NeurIPS 2025 · similarity 0.86)
- Reasoning in Visual Navigation of End-to-end Trained Agents: A Dynamical Systems Approach (CVPR 2025 · similarity 0.85)
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.
