Research map / 3D vision / 3D generation
Objects as Audio-Visual Modal Sound Fields
Zisen Shao, Zihao Wei, Derong Jin, Ruohan Gao
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 DINO visual features (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | CDPAM: 4.20e-4 -> 2.00e-4 | Improved |
| add modal parameter initialization warm-up (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | L1: 0.045 -> 0.019 | Improved |
| add symmetry-aware feature alignment (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | L1-log: 1.002 -> 0.927 | Improved |
| add residual noise component (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | L1-log: 5.764 -> 0.927 | Improved |
| add residual noise component (ablation) | AV-MSF · Contact localization (ablation subset) | localization error: 43.8% -> 38.4% | Improved |
| use FPS instead of closest-point sampling of training locations (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | L1-log: 1.010 -> 0.927 | No clear effect |
| use 71 views instead of 12 for 3DGS prior (ablation) | AV-MSF · ObjectFolder Real (10-object ablation subset) | L1-log: 1.002 -> 0.927 | Improved |
| replace physics-based DiffSound with AV-MSF | AV-MSF · Contact localization | RMED: 41.78% -> 34.61% | Improved |
| hierarchical frequency parameterization vs unconstrained Audio-SDS (ablation) | AV-MSF · Sound editing | UMAP distance: 4.234 -> 2.753 | Improved |
| compare with KNN on non-symmetric objects | AV-MSF · ObjectFolder Real (16 non-symmetric objects) | L1: 0.0135 -> 0.0110 | Improved |
| compare with KNN on full ObjectFolder Real | AV-MSF · ObjectFolder Real (20% train) | L1-log: 0.930 (KNN) vs 0.951 | Mixed |
| reduce shots from 16 to 1 (ablation) | AV-MSF · ObjectFolder Real (10-object subset) | L1: 0.0143 -> 0.0256 | Hurt |
Closest papers by meaning
- AV-GS: Learning Material and Geometry Aware Priors for Novel View Acoustic Synthesis (NeurIPS 2024 · similarity 0.88)
- Improving Audio-Visual Segmentation with Bidirectional Generation (AAAI 2024 · similarity 0.86)
- Real Acoustic Fields: An Audio-Visual Room Acoustics Dataset and Benchmark (CVPR 2024 · similarity 0.86)
- Acoustic Volume Rendering for Neural Impulse Response Fields (NeurIPS 2024 · similarity 0.86)
- Resounding Acoustic Fields with Reciprocity (NeurIPS 2025 · similarity 0.85)
- Robust Audio-Visual Segmentation via Audio-Guided Visual Convergent Alignment (CVPR 2025 · 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.
