Research map / Point clouds / Shape and surface completion
4Diff: 3D-Aware Diffusion Model for Third-to-First Viewpoint Translation
Feng Cheng, Mi Luo, Huiyu Wang, Alex Dimakis, Lorenzo Torresani, Gedas Bertasius, Kristen Grauman
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 3D-aware rotary cross-attention (ablation) | 4Diff · Ego-Exo4D-VT test | LPIPS: 0.373 -> 0.349 | Improved |
| add egocentric point cloud rasterization prior (ablation) | 4Diff · Ego-Exo4D-VT test | LPIPS: 0.412 -> 0.373 | Improved |
| use pretrained MiDaS depth estimator (ablation) | 4Diff · Ego-Exo4D-VT test | LPIPS: 0.392 -> 0.349 | Improved |
| use DINOv2 feature encoder instead of CLIP (ablation) | 4Diff · Ego-Exo4D-VT test | LPIPS: 0.366 -> 0.349 | Improved |
| replace geometry-free diffusion (3DiM) with 4Diff geometry priors | 4Diff · Ego-Exo4D-VT test | LPIPS: 0.385 -> 0.349 | Improved |
| evaluate on unseen scenes instead of seen scenes | 4Diff · Ego-Exo4D-VT unseen scenes | PSNR / SSIM: PSNR 14.45 vs 3DiM 14.90; SSIM 0.508 vs 0.542 | Mixed |
| use NeRF-based GNT for geometry instead of diffusion models | GNT baseline · Ego-Exo4D-VT test | LPIPS: 0.482 vs 0.349 for 4Diff | Hurt |
Closest papers by meaning
- Efficient-3Dim: Learning a Generalizable Single-image Novel-view Synthesizer in One Day (ICLR 2024 · similarity 0.89)
- Enhancing Novel View Synthesis via Geometry Grounded Set Diffusion (NeurIPS 2026 · similarity 0.88)
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.88)
- Era3D: High-Resolution Multiview Diffusion using Efficient Row-wise Attention (NeurIPS 2024 · similarity 0.88)
- WildFusion: Learning 3D-Aware Latent Diffusion Models in View Space (ICLR 2024 · similarity 0.88)
- EgoDTM: Towards 3D-Aware Egocentric Video-Language Pretraining (NeurIPS 2025 · 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.
