Research map / Video / Video generation
Arbitrary Generative Video Interpolation
Guozhen Zhang, Haiguang Wang, Chunyu Wang, Yuan Zhou, Qinglin Lu, Limin 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 |
|---|---|---|---|
| use TaRoPE instead of MLP-based timestamp injection (ablation) | ArbInterp · MultiInterpBench 32x | FID: 35.2 -> 33.7 | Improved |
| use TaRoPE instead of MLP-based timestamp injection (ablation) | ArbInterp · MultiInterpBench 32x | FVD: 454.2 -> 401.6 | Improved |
| add timestamp injection (vanilla) vs none (ablation) | ArbInterp · MultiInterpBench 32x | FID: 38.7 -> 35.2 | Improved |
| add appearance-motion decoupled conditioning on top of TaRoPE (ablation) | ArbInterp · MultiInterpBench 32x | FID: 33.7 -> 26.5 | Improved |
| add motion tokens (vs w/o motion) (ablation) | ArbInterp · MultiInterpBench 32x | FID: 29.8 -> 26.5 | Improved |
| add prefix-frame appearance conditioning (vs w/o appearance) (ablation) | ArbInterp · MultiInterpBench 32x | FID: 31.7 -> 26.5 | Improved |
| use decoupled conditioning instead of latent concatenation (ablation) | ArbInterp · MultiInterpBench 32x | step time (s): 4.4 -> 2.5 | Improved |
| use decoupled conditioning instead of latent concatenation (ablation) | ArbInterp · MultiInterpBench 32x | FID: 27.6 -> 26.5 | Improved |
| use decoupled conditioning instead of cross-attention conditioning (ablation) | ArbInterp · MultiInterpBench 32x | FID: 28.4 -> 26.5 | Improved |
| use three-stage training instead of stage 3 only (ablation) | ArbInterp · MultiInterpBench 32x | FVD: 357.2 -> 319.9 | Improved |
| use ArbInterp over best baseline (DynamiCrafter) at 2x | DynamiCrafter · MultiInterpBench 2x | FID: 83.6 -> 44.9 | Improved |
| use ArbInterp over DynamiCrafter at 256x (beyond training granularity) | DynamiCrafter · 256x interpolation | FVD: 553.1 -> 242.3 | 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.
- LDMVFI — LDMVFI: Video Frame Interpolation with Latent Diffusion Models (AAAI 2024) · better on 3 of 3 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
- Generative Inbetweening through Frame-wise Conditions-Driven Video Generation (CVPR 2025 · similarity 0.88)
- Disentangled Motion Modeling for Video Frame Interpolation (AAAI 2025 · similarity 0.88)
- Framer: Interactive Frame Interpolation (ICLR 2025 · similarity 0.88)
- Perception-Oriented Video Frame Interpolation via Asymmetric Blending (CVPR 2024 · similarity 0.87)
- EDEN: Enhanced Diffusion for High-quality Large-motion Video Frame Interpolation (CVPR 2025 · similarity 0.87)
- Generative Inbetweening: Adapting Image-to-Video Models for Keyframe Interpolation (ICLR 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.
