Research map / 3D vision / NeRF and radiance fields
Reference-guided Controllable Inpainting of Neural Radiance Fields
Ashkan Mirzaei, Tristan Aumentado-Armstrong, Marcus A. Brubaker, Jonathan Kelly, Alex Levinshtein, Konstantinos G. Derpanis, Igor Gilitschenski
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 masked depth loss from aligned monocular disparity (ablation) | Ours-SD · SPIn-NeRF dataset (FR) | FID: 181.20 -> 116.24 | Improved |
| add view-substitution VDE handling and disocclusion loss (ablation) | Ours-SD · SPIn-NeRF dataset (FR) | FID: 145.60 -> 116.24 | Improved |
| add disocclusion loss (ablation) | Ours-SD · SPIn-NeRF dataset (FR) | FID: 126.74 -> 116.24 | Improved |
| use Stable Diffusion instead of LaMa for the reference inpainting (ablation) | Ours · SPIn-NeRF dataset (FR) | FID: 133.27 -> 116.24 | Improved |
| use a real photo as reference instead of an inpainted view (ablation) | Ours · SPIn-NeRF dataset (FR) | FID: 116.24 -> 104.10 | Improved |
| single-reference inpainting instead of SPIn-NeRF perceptual loss over multiple views | NeRF inpainting · SPIn-NeRF dataset (FR) | FID: 156.64 -> 116.24 | Improved |
| use single inpainted reference rather than multiple in NeRF-In | NeRF-In · SPIn-NeRF dataset (FR) | FID: 238.33 -> 183.23 | Improved |
| add DreamFusion SDS prior to Masked NeRF | Masked NeRF · SPIn-NeRF dataset (FR) | FID: 294.69 -> 264.71 | Improved |
| our method vs SPIn-NeRF-LaMa on no-reference sharpness | Ours-LaMa · SPIn-NeRF videos (NR) | Sharpness: 354.31 -> 394.55 | Improved |
Closest papers by meaning
- Removing Objects From Neural Radiance Fields (CVPR 2023 · similarity 0.93)
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.89)
- InNeRF360: Text-Guided 3D-Consistent Object Inpainting on 360-degree Neural Radiance Fields (CVPR 2024 · similarity 0.89)
- Few-Shot Neural Radiance Fields under Unconstrained Illumination (AAAI 2024 · similarity 0.89)
- SIGNeRF: Scene Integrated Generation for Neural Radiance Fields (CVPR 2024 · similarity 0.89)
- NeRF On-the-go: Exploiting Uncertainty for Distractor-free NeRFs in the Wild (CVPR 2024 · similarity 0.89)
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.
