Research map / 3D vision / Face and shape
RealFusion: 360deg Reconstruction of Any Object From a Single Image
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 |
|---|---|---|---|
| Use RealFusion instead of Shelf-Supervised | Single-image object reconstruction · CO3D mean | F-score: 8.24 -> 9.58 | Helped |
| Use RealFusion instead of Shelf-Supervised | Single-image object reconstruction · CO3D mean | CLIP-similarity: 0.70 -> 0.74 | Helped |
| Use RealFusion instead of Shelf-Supervised | Single-image object reconstruction · CO3D Skateboard | F-score: 7.74 -> 5.89 | Hurt |
| Use RealFusion instead of Shelf-Supervised | Single-image object reconstruction · CO3D Teddybear | F-score: 12.89 -> 10.08 | Hurt |
| Use RealFusion instead of SinNeRF | Single-view novel-view synthesis · NeRF Lego (SinNeRF test views) | PSNR: 21.0 -> 16.5 | Hurt |
| Use RealFusion instead of SinNeRF | Single-view novel-view synthesis · NeRF Lego (SinNeRF test views) | LPIPS: 0.09 -> 0.25 | Hurt |
| Add single-image textual inversion using augmented input images (ablation) | RealFusion · Single-image object reconstruction | reconstruction quality | Helped |
| Add 2D normal-smoothness regularization (ablation) | RealFusion · Single-image object reconstruction | surface quality | Helped |
| Train coarse-to-fine rather than optimizing all grid levels from the start (ablation) | RealFusion · Single-image object reconstruction | surface quality | Helped |
| Use Stable Diffusion rather than a CLIP prior (ablation) | RealFusion · Single-image object reconstruction | reconstruction quality | Helped |
| Use monocular depth loss (ablation) | RealFusion · Single-image object reconstruction | reconstruction quality | Mixed |
| Replace L2 reconstruction loss with LPIPS (ablation) | RealFusion · Single-image object reconstruction | reconstruction quality | No clear effect |
Closest papers by meaning
- ReconFusion: 3D Reconstruction with Diffusion Priors (CVPR 2024 · similarity 0.90)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · similarity 0.89)
- DiffHuman: Probabilistic Photorealistic 3D Reconstruction of Humans (CVPR 2024 · similarity 0.88)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.88)
- Denoising Diffusion via Image-Based Rendering (ICLR 2024 · similarity 0.87)
- The More You See in 2D the More You Perceive in 3D (CVPR 2024 · similarity 0.87)
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.