Research map / 3D vision / Gaussian splatting
3D Gaussian Map with Open-Set Semantic Grouping for Vision-Language Navigation
Jianzhe Gao, Rui Liu, Wenguan Wang
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 Egocentric Scene Map (3D Gaussians from pseudo-lidar) to the baseline (ablation) | DUET-style baseline · R2R val unseen | SR: 72 -> 73 | Improved |
| add Open-Set Semantic Grouping on top of ESM (ablation) | ESM · R2R val unseen | SR: 73 -> 75 | Improved |
| add OSG to ESM + MAP (ablation) | full method · REVERIE val unseen | RGS: 35.20 -> 36.73 | Improved |
| add Multi-Level Action Prediction to ESM + OSG (ablation) | ESM + OSG · R2R val unseen | SR: 75 -> 77 | Improved |
| add MAP to ESM + OSG (ablation) | ESM + OSG · REVERIE val unseen | RGS: 34.83 -> 36.73 | Improved |
| add MAP to ESM without OSG (ablation) | ESM · R2R val unseen | SR: 73 -> 73 | No clear effect |
| add ESM, OSG and MAP together vs the baseline (ablation) | baseline · REVERIE val unseen | RGS: 32.15 -> 36.73 | Improved |
| use instance-level scores alone in MAP (ablation) | MAP · R2R val unseen | SR: 72 -> 74 | Improved |
| use scene level alone in MAP (ablation) | MAP · R2R val unseen | SR: 72 -> 73 | Improved |
| use view level alone in MAP (ablation) | MAP · REVERIE val unseen | RGSPL: 23.03 -> 22.76 | Hurt |
| combine all three MAP levels instead of two (ablation) | MAP · R2R val unseen | SR: 75 -> 77 | Improved |
| use the 3D Gaussian Map method vs BEVBert | VLN agent · REVERIE val unseen | RGS: 34.71 -> 36.73 | Improved |
Closest papers by meaning
- Language Embedded 3D Gaussians for Open-Vocabulary Scene Understanding (CVPR 2024 · similarity 0.90)
- GA-VLN: Geometry-Aware BEV Representation for Efficient Vision-Language Navigation (CVPR 2026 · similarity 0.88)
- GaussTR: Foundation Model-Aligned Gaussian Transformer for Self-Supervised 3D Spatial Understanding (CVPR 2025 · similarity 0.88)
- Tackling View-Dependent Semantics in 3D Language Gaussian Splatting (ICML 2025 · similarity 0.87)
- Volumetric Environment Representation for Vision-Language Navigation (CVPR 2024 · similarity 0.87)
- Hierarchical Semantic-Augmented Navigation: Optimal Transport and Graph-Driven Reasoning for Vision-Language Navigation (NeurIPS 2025 · 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.
