Research map / Video / Video generation
Event-Driven Video Generation
Chika Maduabuchi, Jindong Wang
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 EVD event gating and event losses | DiT-4B · EVD-Bench | VBench Dynamics: 78.9 -> 94.8 | Improved |
| add EVD event gating and event losses | DiT-30B · EVD-Bench | VBench Dynamics: 88.7 -> 95.7 | Improved |
| add EVD | DiT-4B · EVD-Bench | VBench Appearance: 75.4 -> 76.2 | No clear effect |
| add EVD | DiT-30B · EVD-Bench | VBench Appearance: 73.8 -> 78.1 | No clear effect |
| add EVD | DiT-4B · EVD-Bench | human Dynamics preference: 80.3 -> 96.4 | Improved |
| add event realization loss (ablation) | DiT-4B+EVD · EVD-Bench | VBench Dynamics: 91.2 -> 94.8 | Improved |
| add event consistency loss (ablation) | DiT-4B+EVD · EVD-Bench | VBench Dynamics: 92.1 -> 94.8 | Improved |
| add event gating at sampling time on top of training (ablation) | DiT-4B+EVD · EVD-Bench | VBench Dynamics: 90.5 -> 94.8 | Improved |
| learn event signal instead of external motion mask (ablation) | DiT-4B+EVD · EVD-Bench | VBench Dynamics: 84.0 -> 94.8 | Improved |
| anneal the gate instead of constant gate 1.0 (ablation) | DiT-4B+EVD · EVD-Bench | human preference for EVD (Dynamics): 60.5 -> 96.4 | Improved |
| anneal the gate instead of constant gate 0.5 (ablation) | DiT-4B+EVD · EVD-Bench | human preference for EVD (Dynamics): 65.8 -> 96.4 | Improved |
| constant gating instead of annealing (ablation) | DiT-4B+EVD · EVD-Bench | VBench Dynamics: 94.1 -> 94.8 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CogVideo2B — CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformers (ICLR 2023) · better on 4 of 4 shared results
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
- EvDiff: High Quality Video with an Event Camera (ECCV 2026 · similarity 0.89)
- Auto-Regressive Diffusion for Generating 3D Human-Object Interactions (AAAI 2025 · similarity 0.86)
- AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusion (CVPR 2025 · similarity 0.86)
- EIDT-V: Exploiting Intersections in Diffusion Trajectories for Model-Agnostic, Zero-Shot, Training-Free Text-to-Video Generation (CVPR 2025 · similarity 0.86)
- InterDyn: Controllable Interactive Dynamics with Video Diffusion Models (CVPR 2025 · similarity 0.86)
- HOIAnimator: Generating Text-prompt Human-object Animations using Novel Perceptive Diffusion Models (CVPR 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.
