Research map / 3D vision / NeRF and radiance fields
Blending-NeRF: Text-Driven Localized Editing in Neural Radiance Fields
Hyeonseop Song, Seokhun Choi, Hoseok Do, Chul Lee, Taehyeong Kim
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 |
|---|---|---|---|
| use dual-NeRF blending (pretrained plus editable NeRF) instead of fine-tuning a single NeRF | Blending-NeRF · Realistic Synthetic 360 | MP_CLIP: .080 -> .121 | Improved |
| Blending-NeRF vs CLIP-NeRF-c on appearance preservation | Blending-NeRF · Realistic Synthetic 360 | D_L1: .029 -> .051 | Hurt |
| add directional CLIP loss (ablation) | Blending-NeRF · Realistic Synthetic 360 | MP_CLIP: .103 -> .121 | Improved |
| add global CLIP loss (ablation) | Blending-NeRF · Realistic Synthetic 360 | MP_CLIP: .115 -> .121 | Improved |
| add region loss (ablation) | Blending-NeRF · Realistic Synthetic 360 | MP_CLIP: .115 -> .121 | Improved |
| add opacity loss (ablation) | Blending-NeRF · Realistic Synthetic 360 | D_L1: .085 -> .051 | Improved |
| add binary-entropy regularization loss on added density (ablation) | Blending-NeRF · Realistic Synthetic 360 | MP_CLIP: .119 -> .121 | Improved |
| increase CLIP patch size with Instant-NGP backbone (ablation) | Blending-NeRF (Instant-NGP) · qualitative edits | visual sharpness | Improved |
| apply image augmentations (differential and perspective) (ablation) | Blending-NeRF · qualitative edits | visual quality | Improved |
| add text augmentation (ablation) | Blending-NeRF · qualitative edits | visual quality | No clear effect |
Closest papers by meaning
- LAENeRF: Local Appearance Editing for Neural Radiance Fields (CVPR 2024 · similarity 0.90)
- Language-driven Object Fusion into Neural Radiance Fields with Pose-Conditioned Dataset Updates (CVPR 2024 · similarity 0.89)
- ED-NeRF: Efficient Text-Guided Editing of 3D Scene With Latent Space NeRF (ICLR 2024 · similarity 0.88)
- IReNe: Instant Recoloring of Neural Radiance Fields (CVPR 2024 · similarity 0.88)
- NeRF Analogies: Example-Based Visual Attribute Transfer for NeRFs (CVPR 2024 · similarity 0.88)
- InNeRF360: Text-Guided 3D-Consistent Object Inpainting on 360-degree Neural Radiance Fields (CVPR 2024 · similarity 0.88)
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.
