Research map / 3D vision / 3D generation
Multi-View Transformer for 3D Visual Grounding
Shijia Huang, Yilun Chen, Jiaya Jia, Liwei Wang
Check 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 multi-view modeling (ablation) | MVT · Nr3D | accuracy: 46.2% -> 52.3% | Improved |
| Add language-guided supervision (ablation) | MVT · Nr3D | accuracy: 40.8% -> 46.2% | Improved |
| Add a transformer decoder for object relationships and vision-language fusion (ablation) | MVT baseline · Nr3D | accuracy: 36.9% -> 40.4% | Improved |
| Add random rotation augmentation (ablation) | MVT with decoder · Nr3D View-dep. | accuracy: 38.4% -> 35.2% | Hurt |
| Add random rotation augmentation (ablation) | MVT with decoder · Nr3D overall | accuracy: 40.4% -> 40.8% | Improved |
| Train with 4 views rather than 1 view (ablation) | MVT · Nr3D | accuracy: 46.2% -> 51.6% | Improved |
| Test with 4 rather than 1 view after training with 4 views (ablation) | MVT · Nr3D | accuracy: 51.6% -> 55.1% | Improved |
| Increase training and test views from 4 to 8 (ablation) | MVT · Nr3D | accuracy: 55.1% -> 53.4% | Hurt |
| Aggregate views after multi-modal fusion rather than after positional encoding (ablation) | MVT · Nr3D | accuracy: 42.5% -> 55.1% | Improved |
| Use average instead of max aggregation (ablation) | MVT · Nr3D | accuracy: 52.7% -> 55.1% | Improved |
| Use average plus max instead of average aggregation (ablation) | MVT · Nr3D | accuracy: 55.1% -> 55.2% | No clear effect |
| Train with additional Sr3D examples (ablation) | MVT · Nr3D | accuracy: 55.1% -> 58.5% | Improved |
Closest papers by meaning
- MiKASA: Multi-Key-Anchor & Scene-Aware Transformer for 3D Visual Grounding (CVPR 2024 · similarity 0.91)
- From Objects to Anywhere: A Holistic Benchmark for Multi-level Visual Grounding in 3D Scenes (NeurIPS 2025 · similarity 0.91)
- SeeGround: See and Ground for Zero-Shot Open-Vocabulary 3D Visual Grounding (CVPR 2025 · similarity 0.91)
- Multi-Object 3D Grounding with Dynamic Modules and Language-Informed Spatial Attention (NeurIPS 2024 · similarity 0.91)
- G^3-LQ: Marrying Hyperbolic Alignment with Explicit Semantic-Geometric Modeling for 3D Visual Grounding (CVPR 2024 · similarity 0.90)
- Towards CLIP-driven Language-free 3D Visual Grounding via 2D-3D Relational Enhancement and Consistency (CVPR 2024 · similarity 0.90)
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.
