Research map / Video / Video generation
MotionStream: Real-Time Video Generation with Interactive Motion Controls
Joonghyuk Shin, Zhengqi Li, Richard Zhang, Jun-Yan Zhu, Jaesik Park, Eli Shechtman, Xun Huang
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 |
|---|---|---|---|
| distill bidirectional teacher into causal student | Wan 2.1-1.3B teacher · DAVIS validation (FPS vs EPE) | FPS: 0.79 -> 16.7 | Improved |
| distill bidirectional teacher into causal student | Wan 2.1-1.3B teacher · DAVIS validation | EPE: 5.35 -> 7.80 | Hurt |
| use sinusoidal PE with track head instead of RGB-VAE track encoding (ablation) | track representation · DAVIS | EPE: 8.57 -> 6.54 | Improved |
| add attention sink (c3s1w1 vs c3s0w1) (ablation) | MotionStream causal student · Sora Extended | LPIPS: 0.501 -> 0.464 | Improved |
| add attention sink (c3s1w1 vs c3s0w1) (ablation) | MotionStream causal student · Sora Extended | EPE: 26.64 -> 25.34 | Improved |
| increase chunk size from 1 to 3 (ablation) | MotionStream causal student · Sora Extended | EPE: 76.21 -> 25.34 | Improved |
| increase chunk size from 1 to 3 (ablation) | MotionStream causal student · Sora Extended | throughput (FPS): 13.26 -> 16.92 | Improved |
| use sink plus single-chunk window instead of Self Forcing sliding window (ablation) | MotionStream causal student · Sora Extended | LPIPS: 0.480 -> 0.464 | Improved |
| enlarge local attention window beyond one chunk (ablation) | MotionStream causal student · Sora extrapolation | quality | Hurt |
| add more than one sink chunk (ablation) | MotionStream causal student · Sora extrapolation | quality | No clear effect |
| swap full VAE for Tiny VAE decoder (ablation) | MotionStream Wan 2.1-1.3B · Sora demo subset | throughput (FPS): 16.7 -> 29.5 | Improved |
| swap full VAE for Tiny VAE decoder (ablation) | MotionStream Wan 2.1-1.3B · Sora demo subset | LPIPS: 0.360 -> 0.365 | No clear effect |
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.
- Image Conductor — Image Conductor: Precision Control for Interactive Video Synthesis (AAAI 2025) · 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
- Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation (CVPR 2026 · similarity 0.91)
- From Slow Bidirectional to Fast Autoregressive Video Diffusion Models (CVPR 2025 · similarity 0.89)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.89)
- AdaState: Self-Evolving Anchors for Streaming Video Generation (NeurIPS 2026 · similarity 0.88)
- Snap Video: Scaled Spatiotemporal Transformers for Text-to-Video Synthesis (CVPR 2024 · similarity 0.88)
- LAMP: Learn A Motion Pattern for Few-Shot Video Generation (CVPR 2024 · 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.
