Research map / Video / Video generation
Be-Your-Outpainter: Mastering Video Outpainting through Input-Specific Adaptation
Fu-Yun Wang, Xiaoshi Wu, Zhaoyang Huang, Xiaoyu Shi, Dazhong Shen, Guanglu Song, Yu Liu, Hongsheng 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 input-specific adaptation (pseudo-outpainting on the source video) versus ControlNet without adaptation (ablation) | MOTIA · DAVIS | FVD: 515.4 -> 286.3 | Improved |
| train adaptation with random masks (pseudo-outpainting) instead of directly fitting the original video (ablation) | MOTIA · DAVIS | FVD: 312.1 -> 286.3 | Improved |
| add spatial-aware insertion of adapters (ablation) | MOTIA · YouTube-VOS | FVD: 66.52 -> 58.40 | Improved |
| add noise regret on top of spatial-aware insertion (ablation) | MOTIA · YouTube-VOS | PSNR: PSNR 19.97 -> 20.25, FVD 58.40 -> 58.99 | Mixed |
| use MOTIA instead of M3DDM | video outpainting · YouTube-VOS | FVD: 66.62 -> 58.99 | Improved |
| use MOTIA instead of M3DDM on PSNR | video outpainting · DAVIS | PSNR: 20.26 -> 20.36 | No clear effect |
| apply pretrained temporal modules directly to outpainting without adaptation (ablation) | Stable Diffusion · DAVIS | FVD: 672.7 vs M3DDM 300.0 | Hurt |
Closest papers by meaning
- Infinite-Canvas: Higher-Resolution Video Outpainting with Extensive Content Generation (AAAI 2025 · similarity 0.91)
- Unboxed: Geometrically and Temporally Consistent Video Outpainting (CVPR 2025 · similarity 0.90)
- Continuous-Multiple Image Outpainting in One-Step via Positional Query and A Diffusion-based Approach (ICLR 2024 · similarity 0.88)
- Video Diffusion Models Are Strong Video Inpainter (AAAI 2025 · similarity 0.88)
- Keyframe-Guided Creative Video Inpainting (CVPR 2025 · similarity 0.87)
- Generative Omnimatte: Learning to Decompose Video into Layers (CVPR 2025 · similarity 0.87)
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.
