Research map / 3D vision / 3D generation
Context-Nav: Context-Driven Exploration and Viewpoint-Aware 3D Spatial Reasoning for Instance Navigation
Won Shik Jang, Ue-Hwan Kim
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 value-map frontier ranking instead of nearest-frontier exploration (ablation) | Context-Nav · CoIN-Bench Val Seen Synonyms | SR / SPL: -9.7 SR / -6.3 SPL | Improved |
| add VLM category verification for open-set classes (ablation) | Context-Nav · CoIN-Bench Val Seen Synonyms | SR / SPL: -9.2 / -3.8 | Improved |
| add intrinsic attribute checks (ablation) | Context-Nav · CoIN-Bench Val Seen Synonyms | SR / SPL: -7.8 / -3.2 | Improved |
| add viewpoint-aware 3D context relation verification (ablation) | Context-Nav · CoIN-Bench Val Seen Synonyms | SR / SPL: -8.3 / -2.5 | Improved |
| use full caption instead of category-only prompt (ablation) | GOAL-CLIP value map · CoIN-Bench Val Seen Synonyms | SR / SPL: +6.6 SR / +3.3 SPL | Improved |
| use full caption instead of category plus intrinsic prompt (ablation) | GOAL-CLIP value map · CoIN-Bench Val Seen Synonyms | SR / SPL: +3.6 / +1.2 | Improved |
| add intrinsic attributes to category-only prompt (ablation) | GOAL-CLIP value map · CoIN-Bench Val Seen Synonyms | SR / SPL: 13.7 -> 16.7 SR; 7.6 -> 9.7 SPL | Improved |
| use full caption instead of category-only prompt (ablation) | BLIP-2 value map · CoIN-Bench Val Seen Synonyms | SPL: +2.2 SPL; SR 15.9 -> 16.4 | Mixed |
| use full caption instead of intrinsic-attribute prompt (ablation) | BLIP-2 value map · CoIN-Bench Val Seen Synonyms | SR / SPL: 17.8 -> 16.4 SR; +1.3 SPL | Mixed |
| use GOAL-CLIP instead of BLIP-2 with full text prompt (ablation) | value map backbone · CoIN-Bench Val Seen Synonyms | SR / SPL: 16.4 -> 20.3 SR; 9.5 -> 10.9 SPL | Improved |
Closest papers by meaning
- LangMap: A Human-Verified Benchmark for Hierarchical Open-Vocabulary Goal Navigation (NeurIPS 2026 · similarity 0.88)
- NavGPT: Explicit Reasoning in Vision-and-Language Navigation with Large Language Models (AAAI 2024 · similarity 0.87)
- EagleVision: A Dual-Stage Framework with BEV-grounding-based Chain-of-Thought for Spatial Intelligence (CVPR 2026 · similarity 0.87)
- Instance-aware Exploration-Verification-Exploitation for Instance ImageGoal Navigation (CVPR 2024 · similarity 0.87)
- Imagine Before Go: Self-Supervised Generative Map for Object Goal Navigation (CVPR 2024 · similarity 0.87)
- From Correspondence to Actions: Human-Like Multi-Image Spatial Reasoning in Multi-modal Large Language Models (ICML 2026 · 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.
