Research map / Video / Video generation
Trajectory-Stabilized Inference for Diffusion-Based Video Inpainting
Zhanhe Zhang, Jiahua Li, Xu Yang, Kun Wei, Cheng Deng
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 trajectory stabilization | SDI · DAVIS | PSNR: 30.48 -> 32.21 | Improved |
| add trajectory stabilization | BrushNet · DAVIS | LPIPS: 0.0478 -> 0.0248 | Improved |
| add trajectory stabilization | DiffuEraser · HQVI | VFID: 0.2962 -> 0.2585 | Improved |
| add trajectory stabilization | CoCoCo · VPBench-L | VFID: 0.62 -> 0.46 | Improved |
| add trajectory stabilization | BrushNet · HQVI | SSIM: 0.9436 -> 0.9524 | No clear effect |
| add trajectory stabilization | SDI · Speed (s/frame) | runtime: 0.965 -> 0.797 | Improved |
| add adaptive anchor selection over uniform sampling (ablation) | SDI + Ours · DAVIS | PSNR: 31.72 -> 32.21 | Improved |
| add anchor-level stabilization (ablation) | SDI + Ours · DAVIS | VFID: 0.191 -> 0.128 | Improved |
| add neighborhood-level stabilization (ablation) | SDI + Ours · DAVIS | VFID: 0.145 -> 0.128 | Improved |
| add transition blending (ablation) | SDI + Ours · DAVIS | VFID: 0.140 -> 0.128 | Improved |
| use risk-gated trajectory stabilization over propagation-based ProPainter on long videos | SDI + Ours · VPBench-L | PSNR: 20.11 -> 22.95 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 on most shared results.
- CoCoCo — CoCoCo: Improving Text-Guided Video Inpainting for Better Consistency, Controllability and Compatibility (AAAI 2025) · better on 4 of 4 shared results
- ProPainter — ProPainter: Improving Propagation and Transformer for Video Inpainting (ICCV 2023) · better on 1 of 3 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
- Reconstruct, Inpaint, Test-Time Finetune: Dynamic Novel-view Synthesis from Monocular Videos (NeurIPS 2025 · similarity 0.87)
- HomoGen: Enhanced Video Inpainting via Homography Propagation and Diffusion (CVPR 2025 · similarity 0.87)
- Tracing the Roots: Leveraging Temporal Dynamics in Diffusion Trajectories for Origin Attribution (NeurIPS 2025 · similarity 0.87)
- Video Diffusion Models Are Strong Video Inpainter (AAAI 2025 · similarity 0.87)
- Generative Inbetweening: Adapting Image-to-Video Models for Keyframe Interpolation (ICLR 2025 · similarity 0.87)
- Motion Consistency Model: Accelerating Video Diffusion with Disentangled Motion-Appearance Distillation (NeurIPS 2024 · 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.
