Research map / Video / Video generation
FramePainter: Endowing Interactive Image Editing with Video Diffusion Priors
Yabo Zhang, Xinpeng Zhou, Yihan Zeng, Hang Xu, Hui Li, Wangmeng Zuo
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 |
|---|---|---|---|
| use matching attention instead of vanilla temporal attention (ablation) | FramePainter · Sketch editing | LPIPS: 0.165 -> 0.140 | Improved |
| use matching attention instead of cross-frame attention (ablation) | FramePainter · Sketch editing | SSIM: 0.826 -> 0.859 | Improved |
| use matching attention instead of cross-frame attention (ablation) | FramePainter · Sketch editing | CLIP-FID: 8.099 -> 7.783 | Improved |
| also reconstruct the source image in the diffusion loss (ablation) | FramePainter · Sketch editing | SSIM: 0.834 -> 0.859 | Improved |
| also reconstruct the source image in the diffusion loss (ablation) | FramePainter · Sketch editing | CLIP-FID: 8.201 -> 7.783 | Improved |
| use FramePainter (20k samples) instead of Magic Fixup (2,500k samples) for coarse edit | SVD-based editor · Coarse edit | SSIM: 0.855 -> 0.888 | Improved |
| use FramePainter instead of LightningDrag (220k samples) for point dragging | SVD-based editor · Point-drag editing | CLIP-FID: 9.894 -> 8.513 | Improved |
| use FramePainter instead of MasaCtrl + ControlNet for sketch editing | SVD-based editor · Sketch editing | CLIP-FID: 17.933 -> 7.783 | Improved |
Closest papers by meaning
- FADE: Frequency-Aware Diffusion Model Factorization for Video Editing (CVPR 2025 · similarity 0.91)
- VINCIE-NExT: Unlocking Video Editing from Images via In-Context Modeling (NeurIPS 2026 · similarity 0.89)
- MotionEditor: Editing Video Motion via Content-Aware Diffusion (CVPR 2024 · similarity 0.89)
- PS-Diffusion: Photorealistic Subject-Driven Image Editing with Disentangled Control and Attention (CVPR 2025 · similarity 0.89)
- AVID: Any-Length Video Inpainting with Diffusion Model (CVPR 2024 · similarity 0.89)
- Keyframe-Guided Creative Video Inpainting (CVPR 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.
