Research map / Video / Video generation
VideoComposer: Compositional Video Synthesis with Motion Controllability
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 |
|---|---|---|---|
| add motion vectors as a condition | VideoComposer without STC-encoder · Motion control evaluation | motion control: 4.03 -> 2.67 | Helped |
| add STC-encoder (ablation) | VideoComposer with text and motion vectors · Motion control evaluation | motion control: 2.67 -> 2.18 | Helped |
| add STC-encoder (ablation) | VideoComposer with text and sketch sequence · Text and sketch sequence | frame consistency: 0.910 -> 0.923 | Helped |
| add STC-encoder (ablation) | VideoComposer with text and depth sequence · Text and depth sequence | frame consistency: 0.922 -> 0.928 | Helped |
| add STC-encoder (ablation) | VideoComposer with text and motion vectors · Text and motion vectors | frame consistency: 0.915 -> 0.927 | Helped |
| use compositional VideoComposer instead of first-stage text-to-video pre-training | VideoComposer · MSR-VTT | FVD: 803 -> 580 | Helped |
| use compositional VideoComposer instead of first-stage text-to-video pre-training | VideoComposer · MSR-VTT | CLIPSIM: 0.2876 -> 0.2932 | Helped |
Closest papers by meaning
- SketchVideo: Sketch-based Video Generation and Editing (CVPR 2025 · similarity 0.88)
- CAGE: Unsupervised Visual Composition and Animation for Controllable Video Generation (AAAI 2025 · similarity 0.86)
- Boximator: Generating Rich and Controllable Motions for Video Synthesis (ICML 2024 · similarity 0.86)
- ONE-SHOT: Compositional Human-Environment Video Synthesis via Spatial-Decoupled Motion Injection and Hybrid Context Integration (NeurIPS 2026 · similarity 0.85)
- Encapsulated Composition of Text-to-Image and Text-to-Video Models for High-Quality Video Synthesis (CVPR 2025 · similarity 0.85)
- Enhancing Motion in Text-to-Video Generation with Decomposed Encoding and Conditioning (NeurIPS 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.