Research map / 3D vision / 3D generation
OBJECT 3DIT: Language-guided 3D-aware Image Editing
Oscar Michel, Anand Bhattad, Eli VanderBilt, Ranjay Krishna, Aniruddha Kembhavi, Tanmay Gupta
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 3DIT 1-task instead of chained-foundation-model baseline | 3DIT · Translation, seen objects | PSNR: 13.699 -> 14.546 | Improved |
| use 3DIT 1-task instead of baseline | 3DIT · Rotation, seen objects | PSNR: 13.179 -> 16.828 | Improved |
| use 3DIT 1-task instead of baseline | 3DIT · Insertion, seen objects | PSNR: 12.297 -> 13.469 | Improved |
| use 3DIT 1-task instead of baseline | 3DIT · Removal, seen objects | PSNR: 12.494 -> 24.937 | Improved |
| use 3DIT 1-task instead of baseline | 3DIT · Translation, seen objects | SSIM: 0.309 -> 0.273 | Hurt |
| use 3DIT 1-task instead of baseline | 3DIT · Translation, seen objects | LPIPS: 0.485 -> 0.494 | Hurt |
| use 3DIT 1-task instead of baseline | 3DIT · Rotation, seen objects | FID: 0.997 -> 0.291 | Improved |
| train one multitask model instead of single-task (ablation) | 3DIT · Translation, seen objects | PSNR: 14.546 -> 15.21 | Improved |
| train one multitask model instead of single-task (ablation) | 3DIT · Rotation, seen objects | PSNR: 16.828 -> 16.859 | No clear effect |
| train one multitask model instead of single-task (ablation) | 3DIT · Removal, seen objects | PSNR: 24.937 -> 24.980 | No clear effect |
| train one multitask model instead of single-task (ablation) | 3DIT · Insertion, seen objects | FID: 0.254 -> 0.222 | Improved |
| evaluate on unseen instead of seen object categories | 3DIT 1-task · Rotation | PSNR: 16.828 -> 16.293 | Hurt |
Closest papers by meaning
- EditVerse3D: High-Quality 3D Object Editing with Region-Aware Learning (ECCV 2026 · similarity 0.92)
- Edit3r: Instant 3D Scene Editing from Sparse Unposed Images (ECCV 2026 · similarity 0.90)
- Towards Scalable and Consistent 3D Editing (ICML 2026 · similarity 0.90)
- SHAP-EDITOR: Instruction-Guided Latent 3D Editing in Seconds (CVPR 2024 · similarity 0.89)
- 3DitScene: Editing Any Scene via Language-guided Disentangled Gaussian Splatting (ICLR 2025 · similarity 0.88)
- Thinking in Boxes: 3D Editing in Real Images Made Easy (NeurIPS 2026 · 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.
