Research map / 3D vision / Depth and stereo
NeuralFusion: Online Depth Fusion in Latent Space
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 |
|---|---|---|---|
| increase the latent feature dimension from N=2 to N=4 (ablation) | NeuralFusion · ModelNet, noisy and outlier-contaminated depth maps | IoU: 0.717 -> 0.863 | Helped |
| increase the latent feature dimension from N=4 to N=8 (ablation) | NeuralFusion · ModelNet, noisy and outlier-contaminated depth maps | IoU: 0.863 -> 0.862 | No clear effect |
| increase the latent feature dimension from N=8 to N=16 (ablation) | NeuralFusion · ModelNet, noisy and outlier-contaminated depth maps | IoU: 0.862 -> 0.863 | No clear effect |
| use feature dimension N=1 (ablation) | NeuralFusion · ModelNet, noisy and outlier-contaminated depth maps | convergence | Hurt |
| train on one chair object instead of the full training set (ablation) | NeuralFusion · ModelNet, noisy depth maps | IoU: 0.874 -> 0.848 | Hurt |
| combine the loss terms rather than using an incomplete loss (ablation) | NeuralFusion · synthetic data | reconstruction quality | Helped |
| translate the fused latent representation into TSDF output to filter outliers | NeuralFusion · real-world depth-map reconstruction | outlier removal | Helped |
| preprocess depth maps with the RoutedFusion routing network | NeuralFusion · ModelNet, varying input noise levels | robustness to high input noise | Helped |
Closest papers by meaning
- RoutedFusion: Learning Real-Time Depth Map Fusion (CVPR 2020 · similarity 0.91)
- Learning Non-Volumetric Depth Fusion Using Successive Reprojections (CVPR 2019 · similarity 0.88)
- Improving Depth Completion via Depth Feature Upsampling (CVPR 2024 · similarity 0.86)
- PatchFusion: An End-to-End Tile-Based Framework for High-Resolution Monocular Metric Depth Estimation (CVPR 2024 · similarity 0.86)
- Depth Prompting for Sensor-Agnostic Depth Estimation (CVPR 2024 · similarity 0.85)
- 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 2024 · 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.