Research map / Video / Video generation
CubeComposer: Spatio-Temporal Autoregressive 4K 360deg Video Generation from Perspective Video
Lingen Li, Guangzhi Wang, Xiaoyu Li, Zhaoyang Zhang, Qi Dou, Jinwei Gu, Tianfan Xue, Ying Shan
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 future fragment tokens from the perspective video to the context (ablation) | CubeComposer · ODV360, 27-frame videos | FVD: 6.0369 -> 4.2592 | Improved |
| use selective context instead of full tokens (ablation) | CubeComposer · ODV360, 27-frame videos | TFLOPs: 376.03 -> 350.64 | Improved |
| use selective context instead of full tokens (ablation) | CubeComposer · ODV360, 27-frame videos | FVD: 5.2265 -> 4.2592 | Improved |
| use selective context instead of full tokens (ablation) | CubeComposer · ODV360, 27-frame videos | FID: 116.6476 -> 125.5510 | Hurt |
| add cube-aware padding and blending (ablation) | CubeComposer · ODV360, 50-epoch ablation | FID: 201.4123 -> 157.1220 | Improved |
| add cube-aware positional encoding (ablation) | CubeComposer · ODV360, 50-epoch ablation | FID: 190.3326 -> 157.1220 | Improved |
| native 4K generation instead of VEnhancer super-resolution of 1K outputs | Argus · 4K360Vid | FID: 168.9571 -> 130.9209 | Improved |
| add VEnhancer post-hoc super-resolution to previous methods | Argus · 4K360Vid | FID: 141.1540 -> 168.9571 | Hurt |
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.
- ViewPoint (1K) — ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models (NeurIPS 2025) · better on 2 of 2 shared results
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
- Imagine360: Immersive 360 Video Generation from Perspective Anchor (NeurIPS 2025 · similarity 0.89)
- ViewPoint: Panoramic Video Generation with Pretrained Diffusion Models (NeurIPS 2025 · similarity 0.88)
- DynamicScaler: Seamless and Scalable Video Generation for Panoramic Scenes (CVPR 2025 · similarity 0.87)
- SpatialDreamer: Self-supervised Stereo Video Synthesis from Monocular Input (CVPR 2025 · similarity 0.87)
- MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views (NeurIPS 2024 · similarity 0.87)
- PanSplat: 4K Panorama Synthesis with Feed-Forward Gaussian Splatting (CVPR 2025 · 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.
