Research map / 3D vision / Depth and stereo
CogniMap3D: Cognitive 3D Mapping and Rapid Retrieval
Feiran Wang, Junyi Wu, Dawen Cai, Yuan Hong, Yan Yan
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 |
|---|---|---|---|
| add 3D features to 2D matching (ablation) | CogniMap3D memory recall · DAVIS scene matching | accuracy (%): 93.5 -> 96.8 | Improved |
| add ICP geometric validation (ablation) | CogniMap3D memory recall · DAVIS scene matching | accuracy (%): 96.8 -> 97.5 | Improved |
| add ICP validation and 3D features (ablation) | CogniMap3D memory recall · DAVIS scene matching | throughput (q/s): 18.2 -> 6.8 | Hurt |
| add memory update (ablation) | CogniMap3D · 7-Scenes reconstruction | Acc mean: 0.086 -> 0.082 | Improved |
| add memory update (ablation) | CogniMap3D · 7-Scenes reconstruction | NC mean: 0.751 -> 0.789 | Improved |
| add memory (ablation) | CogniMap3D · Long-sequence KITTI pose | ATE: 0.068 -> 0.060 | Improved |
| use factor graph refinement instead of PnP+RANSAC (ablation) | CogniMap3D · Camera pose analysis (TUM-dynamic) | ATE: 0.025 -> 0.012 | Improved |
| use factor graph refinement instead of DPVO (ablation) | CogniMap3D · Camera pose analysis (TUM-dynamic) | ATE: 0.019 -> 0.012 | Improved |
| add factor graph optimization (ablation) | CogniMap3D · Camera pose analysis (TUM-dynamic) | RPE rot: 0.510 -> 0.311 | Improved |
| add severe noise to initial poses | CogniMap3D · TUM-dynamic | ATE: 0.012 -> 0.019 | Hurt |
| increase memory size from 50 to 200 | CogniMap3D · DAVIS scene matching | accuracy (%): 96 -> 97.5 | No clear effect |
| use CogniMap3D memory instead of VGGT | CogniMap3D · KITTI reconstruction | Acc mean: 0.071 -> 0.052 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 9 of 10 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- 3D-Mem: 3D Scene Memory for Embodied Exploration and Reasoning (CVPR 2025 · similarity 0.88)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.86)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.86)
- Video World Models with Long-term Spatial Memory (NeurIPS 2025 · similarity 0.86)
- VLM-3R: Vision-Language Models Augmented with Instruction-Aligned 3D Reconstruction (CVPR 2026 · similarity 0.85)
- ImageNet3D: Towards General-Purpose Object-Level 3D Understanding (NeurIPS 2024 · 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.
