Research map / Video / Video segmentation and tracking
Flow-Guided Transformer for Video Inpainting
Kaidong Zhang, Jingjing Fu, Dong Liu
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 |
|---|---|---|---|
| replace FGVC flow completion with LAFC (ablation) | FGT · DAVIS (square masks) | PSNR: 32.49 -> 33.41 | Improved |
| replace FGVC frame synthesis with FGT (ablation) | FGVC · DAVIS (square masks) | PSNR: 32.14 -> 32.49 | Improved |
| add local aggregation to flow completion (ablation) | LAFC · DAVIS (square masks) | EPE: 0.546 -> 0.524 | Improved |
| add edge loss to flow completion (ablation) | LAFC · DAVIS (square masks) | EPE: 0.524 -> 0.511 | Improved |
| add global tokens to local window attention (ablation) | FGT spatial · DAVIS (square masks) | PSNR: 31.37 -> 31.62 | Improved |
| add completed flow tokens without reweight (ablation) | FGT spatial · DAVIS (square masks) | PSNR: 31.62 -> 31.54 | Hurt |
| add flow-reweight module with flow tokens (ablation) | FGT spatial · DAVIS (square masks) | PSNR: 31.54 -> 31.87 | Improved |
| use local-only window instead of dual perspective (ablation) | FGT spatial · DAVIS (square masks) | PSNR: 31.37 -> 31.87 | Improved |
| use flow-guided propagation efficiency tradeoff (all-pair attention) (ablation) | FGT · DAVIS (square masks) | FLOPs: 455.91G -> 703.22G | Hurt |
| FGT vs FFM on Youtube-VOS | FGT · Youtube-VOS | PSNR: 33.73 -> 34.53 | Improved |
| FGT vs FGVC on DAVIS square | FGT · DAVIS (square masks) | PSNR: 32.14 -> 33.41 | Improved |
Closest papers by meaning
- Exploiting Optical Flow Guidance for Transformer-Based Video Inpainting (TPAMI 2024 · similarity 0.94)
- Deep Flow-Guided Video Inpainting (CVPR 2019 · similarity 0.90)
- Towards an End-to-End Framework for Flow-Guided Video Inpainting (CVPR 2022 · similarity 0.89)
- ProPainter: Improving Propagation and Transformer for Video Inpainting (ICCV 2023 · similarity 0.88)
- WaveFormer: Wavelet Transformer for Noise-Robust Video Inpainting (AAAI 2024 · similarity 0.88)
- Video Motion Transfer with Diffusion Transformers (CVPR 2025 · similarity 0.85)
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.
