Research map / Video / Video segmentation and tracking
Long-Term Video Frame Interpolation via Feature Propagation
Dawit Mureja Argaw, In So Kweon
Check 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 motion supervision through the M2FNet loss (ablation) | P-INet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 25.09 -> 27.70; 25.16 -> 28.43 | Improved |
| Add inter-frame motion supervision between forecasted features (ablation) | P-INet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 25.81 -> 27.70; 26.11 -> 28.43 | Improved |
| Add direction supervision between the final propagated feature and reference feature (ablation) | P-INet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 27.13 -> 27.70; 27.83 -> 28.43; 0.48 dB | Improved |
| Add gradient difference loss (ablation) | P-INet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 26.97 -> 27.70; 27.84 -> 28.43; 0.65 dB | Improved |
| Add the local motion decoder (ablation) | M2FNet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 25.13 -> 27.70; 25.71 -> 28.43 | Improved |
| Add the global motion decoder (ablation) | M2FNet · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 26.96 -> 27.70; 27.34 -> 28.43; 0.92 dB | Improved |
| Attend to all past features rather than only the anchor feature (ablation) | P-INet frame decoder · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 26.82 -> 27.70; 27.41 -> 28.43 | Improved |
| Include attended past features in frame decoding (ablation) | P-INet frame decoder · Adobe240 and GOPRO 10 fps → 240 fps | PSNR: 26.03 -> 27.70; 26.57 -> 28.43 | Improved |
| Use P-INet instead of DAIN for low-frame-rate interpolation | Video frame interpolation · Adobe240, GOPRO, and NfS at 15 fps and 8 fps | PSNR: +2.44 dB at 15 fps; +3.51 dB at 8 fps | Improved |
| Use P-INet rather than SloMo for 10 fps → 240 fps up-conversion | Multi-frame VFI · Adobe240, GOPRO, and NfS | PSNR performance range: 8.62 dB -> 6.11 dB | Improved |
Closest papers by meaning
- IQ-VFI: Implicit Quadratic Motion Estimation for Video Frame Interpolation (CVPR 2024 · similarity 0.88)
- Repurposing Pre-trained Video Diffusion Models for Event-based Video Interpolation (CVPR 2025 · similarity 0.88)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.88)
- VFIMamba: Video Frame Interpolation with State Space Models (NeurIPS 2024 · similarity 0.87)
- BiM-VFI: Bidirectional Motion Field-Guided Frame Interpolation for Video with Non-uniform Motions (CVPR 2025 · similarity 0.87)
- LDMVFI: Video Frame Interpolation with Latent Diffusion Models (AAAI 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.
