Research map / Video / Video generation
4DSTR: Advancing Generative 4D Gaussians with Spatial-Temporal Rectification for High-Quality and Consistent 4D Generation
Mengmeng Liu, Jiuming Liu, Yunpeng Zhang, Jiangtao Li, Michael Ying Yang, Francesco Nex, Hao Cheng
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 temporal correlation and spatial rectification to the baseline (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 74.24 -> 45.31 | Improved |
| add Mamba-based temporal correlation on top of spatial rectification (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 55.32 -> 45.31 | Improved |
| add per-frame adaptive densification and pruning on top of temporal rectification (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 52.21 -> 45.31 | Improved |
| use Mamba instead of GRU for temporal encoding (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 50.32 -> 45.31 | Improved |
| use Mamba instead of attention for temporal encoding (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 54.23 -> 45.31 | Improved |
| increase temporal buffer from T=2 to T=10 (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 57.32 -> 45.31 | Improved |
| increase temporal buffer from T=10 to T=15 (ablation) | 4DSTR · Video-to-4D ablation | FID-VID: 45.31 -> 43.54 | No clear effect |
| use 4DSTR instead of STAG4D on 60-frame sequences | video-to-4D generation · Video-to-4D, extended 60-frame sequences | FID-VID: 76.00 -> 43.72 | Improved |
| use 4DSTR instead of STAG4D | video-to-4D generation · Video-to-4D (Consistent4D test set) | FVD: 992 -> 795 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- STAG4D — STAG4D: Spatial-Temporal Anchored Generative 4D Gaussians (ECCV 2024) · better on 7 of 7 shared results
- MVTokenFlow — MVTokenFlow: High-quality 4D Content Generation using Multiview Token Flow (ICLR 2025) · better on 2 of 2 shared results
- Consistent4D — Consistent4D: Consistent 360° Dynamic Object Generation from Monocular Video (ICLR 2024) · better on 1 of 1 shared result
- 4Diffusion — 4Diffusion: Multi-view Video Diffusion Model for 4D Generation (NeurIPS 2024) · 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
- STAG4D: Spatial-Temporal Anchored Generative 4D Gaussians (ECCV 2024 · similarity 0.91)
- Align Your Gaussians: Text-to-4D with Dynamic 3D Gaussians and Composed Diffusion Models (CVPR 2024 · similarity 0.91)
- Instant4D: 4D Gaussian Splatting in Minutes (NeurIPS 2025 · similarity 0.90)
- Real-time Photorealistic Dynamic Scene Representation and Rendering with 4D Gaussian Splatting (ICLR 2024 · similarity 0.89)
- 1000+ FPS 4D Gaussian Splatting for Dynamic Scene Rendering (NeurIPS 2025 · similarity 0.89)
- 4D Gaussian Splatting for Real-Time Dynamic Scene Rendering (CVPR 2024 · 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.
