Research map / Video / Video segmentation and tracking
CCVS: Context-aware Controllable Video Synthesis
Guillaume Le Moing, Jean Ponce, Cordelia Schmid
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 flow-based context recovery (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 389_{\pm12} -> 98_{\pm2} | Improved |
| Add flow self-supervision (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 98_{\pm2} -> 87_{\pm3} | Improved |
| Increase the context window from 1 to 8 frames (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 87_{\pm3} -> 62_{\pm1} | Improved |
| Increase the context window from 8 to 15 frames (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 62_{\pm1} -> 60_{\pm1} | Improved |
| Train for three times as many epochs (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 60_{\pm1} -> 45_{\pm1} | Improved |
| Add a VGG feature loss (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 1200_{\pm6} -> 700_{\pm11} | Improved |
| Add an image discriminator (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 700_{\pm11} -> 323_{\pm3} | Improved |
| Add a temporal discriminator (ablation) | CCVS autoencoder · BAIR reconstruction | FVD: 323_{\pm3} -> 389_{\pm12}; 62_{\pm1} -> 60_{\pm1} | Mixed |
| Condition on the end frame | CCVS · BAIR point-to-point synthesis | SSIM (t=15): 0.683 -> 0.839 | Improved |
| Condition on arm state | CCVS · BAIR state-conditioned synthesis | FVD: 80_{\pm3} -> 50_{\pm1} | Improved |
| Condition on layouts | CCVS · Cityscapes layout-conditioned synthesis | SSIM (t=30): 0.582_{\pm0.005} -> 0.640_{\pm0.005} | Improved |
| Replace ResNet blocks with SPADE blocks (ablation) | CCVS layout-conditioned synthesis · Cityscapes | SSIM (t=30): 0.640_{\pm0.005} -> 0.723_{\pm0.005} | Improved |
Later papers that compared with this method
6 later accepted papers report CCVS as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- CanvasMAR: Improving Masked Autoregressive Video Prediction With Canvas (NeurIPS 2026 · better on 2 of 2 shared results)
- Preserve Your Own Correlation: A Noise Prior for Video Diffusion Models (ICCV 2023 · better on 1 of 1 shared result)
- Sound-Guided Semantic Video Generation (ECCV 2022 · better on 6 of 6 shared results)
- Long Video Generation with Time-Agnostic VQGAN and Time-Sensitive Transformer (ECCV 2022 · better on 3 of 3 shared results)
- MCVD - Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation (NeurIPS 2022 · better on 1 of 1 shared result)
- NÜWA: Visual Synthesis Pre-training for Neural visUal World creAtion (ECCV 2022 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Sound2Sight — Sound2Sight: Generating Visual Dynamics from Sound and Context (ECCV 2020) · better on 0 of 1 shared result
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
- Video-to-Video Synthesis (NeurIPS 2018 · similarity 0.87)
- Video In-context Learning: Autoregressive Transformers are Zero-Shot Video Imitators (ICLR 2025 · similarity 0.87)
- Generative Pre-trained Autoregressive Diffusion Transformer (NeurIPS 2025 · similarity 0.85)
- VSTAR: Generative Temporal Nursing for Longer Dynamic Video Synthesis (ICLR 2025 · similarity 0.85)
- VEDIT: Latent Prediction Architecture For Procedural Video Representation Learning (ICLR 2025 · similarity 0.85)
- ONE-SHOT: Compositional Human-Environment Video Synthesis via Spatial-Decoupled Motion Injection and Hybrid Context Integration (NeurIPS 2026 · 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.
