Research map / Video / Video segmentation and tracking
MotionRNN: A Flexible Model for Video Prediction With Spacetime-Varying Motions
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 MotionRNN to PredRNN | PredRNN · Human3.6M | MSE/10: 48.4 -> 34.2 | Helped |
| Add MotionRNN to PredRNN | PredRNN · Shanghai radar | CSI50: 0.581 -> 0.623 | Helped |
| Add MotionRNN to PredRNN | PredRNN · Varied Moving MNIST | MSE: 35.6 -> 25.1 | Helped |
| Add MotionRNN to MIM | MIM · Guangzhou radar | CSI30: 0.429 -> 0.431 | Helped |
| Add Motion Highway only (ablation) | PredRNN · Human3.6M | MSE/10: 48.4 -> 42.5 | Helped |
| Add trending momentum to MotionGRU without Motion Highway (ablation) | PredRNN + MotionGRU · Human3.6M | MSE/10: 43.5 -> 40.3 | Helped |
| Add transient variation to MotionGRU without Motion Highway (ablation) | PredRNN + MotionGRU · Human3.6M | MSE/10: 41.5 -> 40.3 | Helped |
| Add Motion Highway to MotionGRU with both motion components (ablation) | PredRNN + MotionGRU · Human3.6M | MSE/10: 40.3 -> 34.2 | Helped |
| Add trending momentum to MotionRNN (ablation) | PredRNN + MotionRNN · Human3.6M | MSE/10: 38.9 -> 34.2 | Helped |
| Add transient variation to MotionRNN (ablation) | PredRNN + MotionRNN · Human3.6M | MSE/10: 40.6 -> 34.2 | Helped |
| Vary trending-momentum step size alpha (ablation) | MotionRNN with PredRNN and ConvLSTM · Human3.6M and other benchmarks | prediction performance | No clear effect |
Closest papers by meaning
- Motion Graph Unleashed: A Novel Approach to Video Prediction (NeurIPS 2024 · similarity 0.87)
- DynaTokens: Teaching Dynamics to Camera-Controlled Video Models at Test Time (NeurIPS 2026 · similarity 0.85)
- LaMo: Self-Supervised Latent Motion Priors for Physical Realism in Video Generation (NeurIPS 2026 · similarity 0.85)
- E-Motion: Future Motion Simulation via Event Sequence Diffusion (NeurIPS 2024 · similarity 0.84)
- Wan-Move: Motion-controllable Video Generation via Latent Trajectory Guidance (NeurIPS 2025 · similarity 0.84)
- Wavelet-Driven Spatiotemporal Predictive Learning: Bridging Frequency and Time Variations (AAAI 2024 · similarity 0.84)
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.