Research map / Video / Video segmentation and tracking
Probabilistic Future Prediction for Video Scene Understanding
Anthony Hu, Fergal Cotter, Nikhil Mohan, Corina Gurau, Alex Kendall
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 probabilistic present/future distributions | Temporal Block model · Urban driving data | M_perception: 13.6% -> 20.0% | Improved |
| add probabilistic modelling | Res. 3D Conv. · Urban driving data | M_perception: 6.9% -> 8.1% | Improved |
| add probabilistic modelling | Conv. GRU · Urban driving data | M_perception: 7.4% -> 9.0% | Improved |
| add probabilistic modelling | Sep. Inception · Urban driving data | M_perception: 9.6% -> 13.8% | Improved |
| use proposed Temporal Block instead of Sep. Inception | probabilistic model · Urban driving data | M_perception: 13.8% -> 20.0% | Improved |
| add future prediction loss to policy training (ablation) | driving policy · Urban driving data | steering error: 0.043 -> 0.036 | Improved |
| add future prediction loss to policy training (ablation) | driving policy · Urban driving data | speed error: 0.039 -> 0.030 | Improved |
| make future prediction probabilistic | driving policy · Urban driving data | steering error: 0.036 -> 0.033 | Improved |
| make future prediction probabilistic | driving policy · Urban driving data | speed error: 0.030 -> 0.026 | Improved |
| add global context pooling to Temporal Block (ablation) | Temporal Block · Cityscapes (i=5) | IoU: 0.453 -> 0.464 | Improved |
| add global context pooling to Temporal Block (ablation) | Temporal Block · Cityscapes (i=10) | IoU: 0.413 -> 0.416 | Improved |
| add kt,1,ks convolution to Temporal Block without global context (ablation) | Temporal Block · Cityscapes (i=5) | IoU: 0.454 -> 0.461 | Mixed |
Closest papers by meaning
- An Uncertain Future: Forecasting from Static Images Using Variational Autoencoders (ECCV 2016 · similarity 0.92)
- DESIRE: Distant Future Prediction in Dynamic Scenes With Interacting Agents (CVPR 2017 · similarity 0.90)
- Predicting Deeper Into the Future of Semantic Segmentation (ICCV 2017 · similarity 0.90)
- Compositional Video Prediction (ICCV 2019 · similarity 0.89)
- Advancing Semantic Future Prediction through Multimodal Visual Sequence Transformers (CVPR 2025 · similarity 0.88)
- Towards Understanding Future: Consistency Guided Probabilistic Modeling for Action Anticipation (AAAI 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.
