Research map / Video / Video segmentation and tracking
"Clearer Frames, Anytime: Resolving Velocity Ambiguity in Video Frame Interpolation"
Zhihang Zhong, Gurunandan Krishnan, Xiao Sun, Yu Qiao, Sizhuo Ma, Jian Wang
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 time indexing with distance indexing (uniform map at inference) | RIFE · Vimeo90K septuplet | NIQE: 6.663 -> 6.344 | Improved |
| add iterative reference-based estimation on top of distance indexing (ablation) | RIFE · Vimeo90K septuplet | NIQE: 6.344 -> 6.220 | Improved |
| use distance indexing with uniform map at inference | RIFE · Vimeo90K septuplet | PSNR: 28.22 -> 27.55 | Hurt |
| use ground-truth distance maps instead of time indexing | RIFE · Vimeo90K septuplet | PSNR: 28.22 -> 29.20 | Improved |
| increase reference iterations from 2 to 3 (ablation) | distance indexing + reference estimation · Vimeo90K septuplet | LPIPS: 0.086 -> 0.085 (RIFE) | No clear effect |
| add reference-based estimation without distance indexing (ablation) | AMT-S · Vimeo90K septuplet | LPIPS: 0.101 -> 0.135 (2 iterations) | Hurt |
| apply distance indexing and reference estimation at larger temporal distance (x16 vs x8) | RIFE · X4K1000FPS | LPIPS: 0.104 -> 0.079 | Improved |
| use GMFlow instead of RAFT for distance maps (ablation) | RIFE [D,R]u · Vimeo90K septuplet | NIQE: 6.220 -> 6.280 | No clear effect |
| apply distance indexing and reference estimation on easier benchmarks | RIFE · Adobe240 x8 | LPIPS: 0.073 -> 0.054 | Improved |
| use both strategies versus time indexing in user ranking | four VFI models · User study (30 participants) | user ranking | Improved |
Closest papers by meaning
- Velocity Disambiguation for Video Frame Interpolation (TPAMI 2026 · similarity 0.96)
- Video Frame Interpolation without Temporal Priors (NeurIPS 2020 · similarity 0.89)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.87)
- Generative Inbetweening: Adapting Image-to-Video Models for Keyframe Interpolation (ICLR 2025 · similarity 0.86)
- EDEN: Enhanced Diffusion for High-quality Large-motion Video Frame Interpolation (CVPR 2025 · similarity 0.86)
- IQ-VFI: Implicit Quadratic Motion Estimation for Video Frame Interpolation (CVPR 2024 · similarity 0.86)
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.
