Research map / Video / Video segmentation and tracking
Revisiting Hierarchical Approach for Persistent Long-Term Video Prediction
Read the paperCheck 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 |
|---|---|---|---|
| predict dense semantic structures stochastically and translate them into pixels hierarchically | video prediction · Human Dancing, short-term | CSIM: 0.7637 -> 0.8164 | Helped |
| predict dense semantic structures stochastically and translate them into pixels hierarchically | video prediction · Human Dancing, short-term | FVD: 1987.55 -> 1398.98 | Helped |
| predict dense semantic structures stochastically and translate them into pixels hierarchically | video prediction · KITTI, short-term | mIoU: 0.4225 -> 0.5137 | Helped |
| predict dense semantic structures stochastically and translate them into pixels hierarchically | video prediction · KITTI, short-term | FVD: 956.05 -> 762.73 | Helped |
| predict future structures stochastically (ablation) | sequential dense structure generator · KITTI, long-term | shot-wise FVD | Helped |
| use recurrent rather than feedforward structure estimation (ablation) | sequential dense structure generator · KITTI, short- to mid-term (t < 200) | shot-wise FVD | Helped |
| discretize predicted semantic label maps (ablation) | sequential dense structure generator · KITTI, long-term | shot-wise FVD | Helped |
| keep structure prediction independent of generated RGB feedback (ablation) | hierarchical video prediction · KITTI, long-term | shot-wise FVD | Helped |
| add instance-boundary map prediction (ablation) | image generator on Cityscapes · Cityscapes, 6 future frames | PSNR: 19.236 -> 20.427 | Helped |
| add instance-boundary map prediction (ablation) | image generator on Cityscapes · Cityscapes, 6 future frames | CSIM: 0.937 -> 0.938 | No clear effect |
Closest papers by meaning
- Learning to Generate Long-term Future via Hierarchical Prediction (ICML 2017 · similarity 0.93)
- Hierarchical Long-term Video Prediction without Supervision (ICML 2018 · similarity 0.90)
- Video Prediction Recalling Long-Term Motion Context via Memory Alignment Learning (CVPR 2021 · similarity 0.89)
- Video Prediction by Modeling Videos as Continuous Multi-Dimensional Processes (CVPR 2024 · similarity 0.88)
- STDiff: Spatio-Temporal Diffusion for Continuous Stochastic Video Prediction (AAAI 2024 · similarity 0.87)
- Visual Representation Learning with Stochastic Frame Prediction (ICML 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.