Research map / Vision-language and multimodal / Robot action and navigation
Bird's-Eye-View Scene Graph for Vision-Language Navigation
Rui Liu, Xiaohan Wang, Wenguan Wang, Yi Yang
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 BEV scene graph with grid and graph decisions to the basic agent (ablation) | DUET · REVERIE val unseen | SR: 46.98 -> 52.12 | Improved |
| use BEV branch alone without the panoramic branch (ablation) | BSG · REVERIE val unseen | SR: 46.98 -> 39.03 | Hurt |
| add detection loss (ablation) | BSG · REVERIE val unseen | SR: 49.25 -> 52.12 | Improved |
| add BEV updating by cross-attention (ablation) | BSG · REVERIE val unseen | SR: 50.30 -> 52.12 | Improved |
| fuse graph-level and grid-level decision scores (ablation) | BSG · REVERIE val unseen | SR: 50.18 / 48.25 -> 52.12 | Improved |
| use bivariate Gaussian weights instead of uniform grid weights (ablation) | BSG · REVERIE val unseen | SR: 51.27 -> 52.12 | Improved |
| use 9 neighboring grids for node embeddings rather than 4 or 16 (ablation) | BSG · REVERIE val unseen | SR: 51.33 / 51.71 -> 52.12 | Improved |
| use BEVFormer instead of LSS or BEVDepth (ablation) | BSG · REVERIE val unseen | SR: 50.83 / 51.06 -> 52.12 | Improved |
| use oriented bounding boxes instead of axis-aligned boxes (ablation) | BSG · REVERIE val unseen | SR: 49.25 -> 52.12 | Improved |
| use 4 reference points rather than 2 or 8 (ablation) | BSG · REVERIE val unseen | SR: 51.49 / 50.98 -> 52.12 | 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.
- DUET — DUET: Cross-Modal Semantic Grounding for Contrastive Zero-Shot Learning (AAAI 2023) · better on 4 of 4 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
- GA-VLN: Geometry-Aware BEV Representation for Efficient Vision-Language Navigation (CVPR 2026 · similarity 0.91)
- Volumetric Environment Representation for Vision-Language Navigation (CVPR 2024 · similarity 0.88)
- Hierarchical Semantic-Augmented Navigation: Optimal Transport and Graph-Driven Reasoning for Vision-Language Navigation (NeurIPS 2025 · similarity 0.87)
- Improving Bird's Eye View Semantic Segmentation by Task Decomposition (CVPR 2024 · similarity 0.87)
- GPT4Scene: Understand 3D Scenes from Videos with Vision-Language Models (ICLR 2026 · similarity 0.87)
- SG-Nav: Online 3D Scene Graph Prompting for LLM-based Zero-shot Object Navigation (NeurIPS 2024 · similarity 0.86)
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.
