Research map / Video / Video generation
Motion-Residual Conflict-Aware Time Reversal for Generative Inbetweening
Zhenbang Zhang, Zihui Cui, Haythem El-Messiry, Renmin Han, zhiqiang xu
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 |
|---|---|---|---|
| add end-conditioned residual to the consensus motion axis (ablation) | MR-CATR on TRF · DAVIS | FVD: 548.3 -> 523.6 | Improved |
| add end-conditioned residual to the consensus motion axis (ablation) | MR-CATR on ViBiD · DAVIS | FVD: 531.4 -> 519.5 | Improved |
| add residual-guided projection of backward update (ablation) | MR-CATR on TRF · DAVIS | FVD: 552.2 -> 523.6 | Improved |
| add residual-guided projection of backward update (ablation) | MR-CATR on ViBiD · DAVIS | LPIPS: 0.247 -> 0.210 | Improved |
| apply MR-CATR only in the early 0.2T high-noise steps rather than late 0.2T (ablation) | MR-CATR on TRF · DAVIS | FVD: 589.6 -> 467.2 | Improved |
| extend active window from early 0.2T to all steps (ablation) | MR-CATR on TRF · DAVIS | FVD: 467.2 -> 476.5 | No clear effect |
| extend active window from early 0.2T to early 0.5T (ablation) | MR-CATR on ViBiD · DAVIS | FVD: 462.8 -> 466.1 | No clear effect |
| plug MR-CATR into TRF | TRF · DAVIS | FVD: 643.92 -> 523.56 | Improved |
| plug MR-CATR into ViBiD | ViBiD · DAVIS | FID: 41.586 -> 36.930 | Improved |
| use MR-CATR instead of MPD on ViBiD | ViBiD · Pexels | FID: 36.189 -> 37.062 | Hurt |
| use MR-CATR instead of MPD on TRF | TRF · Pexels VBench | motion smoothness: 0.9489 -> 0.9484 | Hurt |
Closest papers by meaning
- Generative Inbetweening: Adapting Image-to-Video Models for Keyframe Interpolation (ICLR 2025 · similarity 0.89)
- AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusion (CVPR 2025 · similarity 0.88)
- TRIP: Temporal Residual Learning with Image Noise Prior for Image-to-Video Diffusion Models (CVPR 2024 · similarity 0.88)
- Inference-Time Text-to-Video Alignment with Diffusion Latent Beam Search (NeurIPS 2025 · similarity 0.88)
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.88)
- Generative Pre-trained Autoregressive Diffusion Transformer (NeurIPS 2025 · similarity 0.88)
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.
