Research map / Video / Video segmentation and tracking
Frame Interpolation Transformer and Uncertainty Guidance
Markus Plack, Karlis Martins Briedis, Abdelaziz Djelouah, Matthias B. Hullin, Markus Gross, Christopher Schroers
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 Ours L1 instead of VFIformer | frame interpolation · SNU-FILM Hard | PSNR: 30.67 -> 30.88 | Improved |
| Use Ours L1 instead of VFIformer | frame interpolation · Vimeo90k | PSNR: 36.50 -> 36.34 | Hurt |
| Add deep feature extraction rather than the Reda et al. feature representation (ablation) | Ours L1 · Vimeo90k | PSNR: 36.28 -> 36.34 | Improved |
| Add deep feature extraction rather than the Reda et al. feature representation (ablation) | Ours L1 · Animated films | PSNR: 35.06 -> 35.75 | Improved |
| Add error estimation outputs and error losses (ablation) | Ours L1 · Vimeo90k | PSNR: 36.31 -> 36.34 | Improved |
| Add error estimation outputs and error losses (ablation) | Ours L1 · Animated films | PSNR: 35.71 -> 35.75 | Improved |
| Process all frames with shared frame updates instead of updating only the center frame (ablation) | Ours L1 · Vimeo90k | PSNR: 35.82 -> 36.34 | Improved |
| Process all frames with shared frame updates instead of updating only the center frame (ablation) | Ours L1 · Animated films | PSNR: 35.28 -> 35.75 | Improved |
| Select tiles using predicted color error rather than random sampling for ground-truth tile replacement | uncertainty-guided interpolation · animated short films | PSNR: 6.99 and 9.98 dB; random replacement yields at most 1.27 dB | Improved |
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.
- IFRNet — IFRNet: Intermediate Feature Refine Network for Efficient Frame Interpolation (CVPR 2022) · 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
- Video Frame Interpolation without Temporal Priors (NeurIPS 2020 · similarity 0.89)
- Super SloMo: High Quality Estimation of Multiple Intermediate Frames for Video Interpolation (CVPR 2018 · similarity 0.89)
- All at Once: Temporally Adaptive Multi-Frame Interpolation with Advanced Motion Modeling (ECCV 2020 · similarity 0.88)
- Perception-Oriented Video Frame Interpolation via Asymmetric Blending (CVPR 2024 · similarity 0.87)
- Video Frame Interpolation With Many-to-Many Splatting and Spatial Selective Refinement (TPAMI 2023 · similarity 0.87)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 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.
