Research map / 3D vision / NeRF and radiance fields
Unsupervised Discovery of Object Radiance Fields
Read the paperCheck 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 |
|---|---|---|---|
| Add background-aware modeling (ablation) | uORF · CLEVR-567 | NV-ARI: 10.5$\pm$3.6 -> \textbf{83.8}$\pm$0.3 | Helped |
| Add background-aware modeling (ablation) | uORF · Room-Chair | LPIPS: 0.1671 -> 0.0821 | Helped |
| Add background-aware modeling (ablation) | uORF · Room-Diverse | NV-ARI: 21.0$\pm$8.1 -> \textbf{56.9}$\pm$0.2 | Helped |
| Add the fine training stage after coarse training (ablation) | uORF · Room-Chair | NV-ARI: 62.3$\pm$2.5 -> \textbf{74.3}$\pm$1.9 | Helped |
| Add the fine training stage after coarse training (ablation) | uORF · Room-Diverse | LPIPS: 0.2123 -> 0.1729 | Helped |
| Add perceptual loss (ablation) | uORF · Room-Diverse | LPIPS: 0.3610 -> 0.2156 | Helped |
| Add adversarial loss (ablation) | uORF with reconstruction and perceptual losses · Room-Diverse | LPIPS: 0.2156 -> 0.1729 | Helped |
| Add perceptual loss (ablation) | uORF with reconstruction and adversarial losses · Room-Diverse | LPIPS: 0.2288 -> 0.1729 | Helped |
| Represent the background in world space rather than viewer space (ablation) | uORF · Room-Chair | LPIPS: 0.151 -> 0.082 | Helped |
| Represent the background in world space rather than viewer space (ablation) | uORF · Room-Chair | NV-ARI: 73.5 -> 74.3 | Helped |
| Use uORF instead of monolithic NeRF-AE | single-image novel-view synthesis · Room-Chair | LPIPS: 0.1166 -> 0.0821 | Helped |
| Use uORF instead of Slot Attention | 3D scene segmentation · Room-Diverse | ARI: 17.4$\pm$11.3 -> \textbf{65.6}$\pm$1.0 | Helped |
Closest papers by meaning
- Slot-guided Volumetric Object Radiance Fields (NeurIPS 2023 · similarity 0.90)
- Unsupervised Multi-View Object Segmentation Using Radiance Field Propagation (NeurIPS 2022 · similarity 0.90)
- Unsupervised Causal Generative Understanding of Images (NeurIPS 2022 · similarity 0.89)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 2024 · similarity 0.88)
- RelationField: Relate Anything in Radiance Fields (CVPR 2025 · similarity 0.87)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · 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.