Research map / 3D vision / 3D generation
AnyView: Synthesizing Any Novel View in Dynamic Scenes
Basile Van Hoorick, Dian Chen, Shun Iwase, Pavel Tokmakov, Muhammad Zubair Irshad, Igor Vasiljevic, Swati Gupta, Fangzhou Cheng, Sergey Zakharov, Vitor Guizilini
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 |
|---|---|---|---|
| initialize from Cosmos pretrained weights instead of training from scratch (ablation) | AnyView · AnyViewBench in-distribution (avg) | PSNR: 16.55 -> 17.61 | Improved |
| initialize from Cosmos pretrained weights instead of training from scratch (ablation) | AnyView · DyCheck iPhone (narrow, zero-shot) | PSNR: 9.40 -> 13.47 | Improved |
| initialize from Cosmos pretrained weights instead of training from scratch (ablation) | AnyView · AnyViewBench zero-shot (avg) | PSNR: 11.35 -> 11.58 | Improved |
| use the full 12-dataset training mixture instead of GCD data only (ablation) | AnyView · AnyViewBench in-distribution (avg) | PSNR: 14.54 -> 17.61 | Improved |
| use the full training mixture instead of GCD data only (ablation) | AnyView · DyCheck iPhone (narrow, zero-shot) | PSNR: 11.38 -> 13.47 | Improved |
| use the full training mixture instead of GCD data only (ablation) | AnyView · ParDom-4D (gradual, narrow) | PSNR: 26.18 vs 26.29 | No clear effect |
| use the full training mixture instead of GCD data only (ablation) | AnyView · Kubric-4D (gradual, narrow) | PSNR: 22.06 -> 21.21 | Hurt |
| implicit Plucker-ray conditioning without depth reprojection vs depth-reprojection baselines | AnyView vs TrajCrafter · AnyViewBench in-distribution (avg) | PSNR: 12.82 -> 17.61 | Improved |
| implicit conditioning vs depth reprojection in narrow DVS | AnyView vs depth/TTO baselines · DyCheck iPhone (narrow, zero-shot) | PSNR: 13.47 vs 14.24 / 16.94 | Mixed |
| compare AnyView with GCD | AnyView vs GCD · AnyViewBench in-distribution (avg) | FVD: 4126 -> 1239 | Improved |
| compare AnyView on estimated camera pose of generated videos | AnyView vs best baseline · AnyViewBench in-distribution (camera pose accuracy) | AUC15: 21.30 -> 43.38 | Improved |
| compare AnyView on zero-shot camera pose accuracy | AnyView vs best baseline · AnyViewBench zero-shot (camera pose accuracy) | AUC15: 12.31 -> 16.52 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GEN3C — GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control (CVPR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Reconstruct, Inpaint, Test-Time Finetune: Dynamic Novel-view Synthesis from Monocular Videos (NeurIPS 2025 · similarity 0.90)
- ViVid-1-to-3: Novel View Synthesis with Video Diffusion Models (CVPR 2024 · similarity 0.89)
- Novel View Synthesis with Pixel-Space Diffusion Models (CVPR 2025 · similarity 0.89)
- NVS-Solver: Video Diffusion Model as Zero-Shot Novel View Synthesizer (ICLR 2025 · similarity 0.89)
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · similarity 0.89)
- UMAMI: Unifying Masked Autoregressive Models and Deterministic Rendering for View Synthesis (NeurIPS 2025 · similarity 0.89)
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.
