Research map / 3D vision / 3D generation
Towards Realistic and Consistent Orbital Video Generation via 3D Foundation Priors
Rong Wang, Ruyi Zha, Ziang Cheng, Jiayu Yang, Pulak Purkait, Hongdong Li
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 global latent vector from 3D foundation model (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | PSNR: 20.06 -> 21.86 | Improved |
| add global latent vector from 3D foundation model (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | MEt3R: 0.08 -> 0.06 | Improved |
| add global latent vector from 3D foundation model (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | CLIP-S: 91.26 -> 93.51 | Improved |
| add local volumetric latent images to global vector (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | PSNR: 21.86 -> 22.78 | Improved |
| add local volumetric latent images to global vector (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | CLIP-S: 93.51 -> 94.19 | Improved |
| add local volumetric latent images to global vector (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | MEt3R: 0.06 -> 0.05 | Improved |
| use cross-attention adapter instead of pooled feature concatenation (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | PSNR: 20.53 -> 22.78 | Improved |
| use cross-attention adapter instead of pooled feature concatenation (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | CLIP-S: 91.88 -> 94.19 | Improved |
| use cross-attention adapter instead of frame stacking (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | PSNR: 19.27 -> 22.78 | Improved |
| use cross-attention adapter instead of frame stacking (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | CLIP-S: 89.96 -> 94.19 | Improved |
| use cross-attention adapter instead of frame stacking (ablation) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | LPIPS: 0.18 -> 0.09 | Improved |
| use 3D foundation priors vs SV3D baseline (full model) | Orbital video model (SVD base) · Objaverse-XL (ablation, 21 frames) | MEt3R: 0.07 -> 0.05 | Improved |
Closest papers by meaning
- Diffusion4D: Fast Spatial-temporal Consistent 4D generation via Video Diffusion Models (NeurIPS 2024 · similarity 0.87)
- GeoVideo: Introducing Geometric Regularization into Video Generation Model (NeurIPS 2025 · similarity 0.87)
- PanoWorld: Geometry-Consistent Panoramic Video World Modeling (NeurIPS 2026 · similarity 0.87)
- Structured 3D Latents for Scalable and Versatile 3D Generation (CVPR 2025 · similarity 0.87)
- World Tracing: Pixel-Aligned Geometry Beyond the Visible (NeurIPS 2026 · similarity 0.86)
- Vidu4D: Single Generated Video to High-Fidelity 4D Reconstruction with Dynamic Gaussian Surfels (NeurIPS 2024 · similarity 0.86)
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.
