Research map / Video / Video segmentation and tracking
Stochastic Image-to-Video Synthesis Using cINNs
Michael Dorkenwald, Timo Milbich, Andreas Blattmann, Robin Rombach, Konstantinos G. Derpanis, Bjorn Ommer
Check 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 |
|---|---|---|---|
| remove the cINN, leaving a cVAE framework (ablation) | Ours · BAIR Robot Pushing | FVD: 99.3 -> 134.5 | Hurt |
| remove input-image conditioning from the generator (ablation) | Ours · BAIR Robot Pushing | FVD: 99.3 -> 272.6 | Hurt |
| remove ADAIN input of z from video residual blocks (ablation) | Ours · BAIR Robot Pushing | FVD: 99.3 -> 131.2 | Hurt |
| remove the cINN, leaving a cVAE framework (ablation) | Ours · iPER | FVD: 132.9 -> 180.6 | Hurt |
| remove input-image conditioning from the generator (ablation) | Ours · iPER | FVD: 132.9 -> 381.5 | Hurt |
| remove ADAIN input of z from video residual blocks (ablation) | Ours · iPER | FVD: 132.9 -> 156.7 | Hurt |
| use the proposed method instead of AL | Image-to-video synthesis · Landscape | LPIPS: 0.26 -> 0.23 | Improved |
| use the proposed method instead of AL | Image-to-video synthesis · Landscape | DIV VGG: 0.97 -> 0.71 | Hurt |
| use the proposed method instead of Video Transformer | Image-to-video synthesis · BAIR Robot Pushing | FVD: 94.0 -> 99.3 | Hurt |
| use the proposed method instead of AL | Image-to-video synthesis · DTDB Clouds | FVD: 142.1 -> 179.3 | Hurt |
Closest papers by meaning
- Video-to-Video Synthesis (NeurIPS 2018 · similarity 0.88)
- Conditional Image-to-Video Generation With Latent Flow Diffusion Models (CVPR 2023 · similarity 0.87)
- Neural 3D Video Synthesis From Multi-View Video (CVPR 2022 · similarity 0.87)
- Decouple Content and Motion for Conditional Image-to-Video Generation (AAAI 2024 · similarity 0.86)
- BIVDiff: A Training-Free Framework for General-Purpose Video Synthesis via Bridging Image and Video Diffusion Models (CVPR 2024 · similarity 0.86)
- TRIP: Temporal Residual Learning with Image Noise Prior for Image-to-Video Diffusion Models (CVPR 2024 · similarity 0.85)
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.
