Research map / 3D vision / Depth and stereo
BNV-Fusion: Dense 3D Reconstruction Using Bi-Level Neural Volume Fusion
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 bi-level fusion instead of TSDF Fusion | BNV-Fusion · 3D Scene Dataset (Lounge, CactusGarden, StoneWall, Burghers) | F1: 89.66 -> 91.01; 83.84 -> 85.75; 91.45 -> 93.37; 77.26 -> 79.08 | Helped |
| Use bi-level fusion instead of TSDF Fusion | BNV-Fusion · CopyRoom | F1: 90.05 -> 89.43 | Hurt |
| Use bi-level fusion instead of the strongest listed baseline | BNV-Fusion · Augmented ICL-NUIM (Livingroom0, Livingroom1, Office0) | F1: 62.55 -> 72.54; 68.00 -> 81.02; 51.41 -> 59.23 | Helped |
| Use bi-level fusion instead of TSDF Fusion | BNV-Fusion · Augmented ICL-NUIM (Office1) | F1: 57.79 -> 63.91 | Helped |
| Use bi-level fusion instead of TSDF Fusion | BNV-Fusion · ScanNet validation | F1: 78.84 -> 80.56 | Helped |
| Add global-level neural-rendering fusion (ablation) | BNV-Fusion · Augmented ICL-NUIM | F1: 34.31 -> 69.17 | Helped |
| Add local-level fusion to initialize global-level fusion (ablation) | BNV-Fusion · Augmented ICL-NUIM | F1: 36.55 -> 69.17 | Helped |
| Initialize global-level fusion with local-level fusion rather than randomly (ablation) | BNV-Fusion · Augmented ICL-NUIM | iterations to converge: 5 iterations; 20 iterations | Helped |
Closest papers by meaning
- Learning Neural Implicit through Volume Rendering with Attentive Depth Fusion Priors (NeurIPS 2023 · similarity 0.89)
- Volume Rendering of Neural Implicit Surfaces (NeurIPS 2021 · similarity 0.87)
- NeuralFusion: Online Depth Fusion in Latent Space (CVPR 2021 · similarity 0.86)
- NeuRodin: A Two-stage Framework for High-Fidelity Neural Surface Reconstruction (NeurIPS 2024 · similarity 0.84)
- 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 2024 · similarity 0.84)
- MVD-Fusion: Single-view 3D via Depth-consistent Multi-view Generation (CVPR 2024 · similarity 0.83)
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.