AI★★★arXiv · 2026-07-23
GraphVid: Interactive Graph-Controllable Video Generation
Controllable video generation is challenging due to difficulties in specifying precise multi-object interactions with text prompts or motion controls. GraphVid addresses this by introducing a graph-conditioned image-to-video generation model that allows for interactive and precise control.
📌 Key points
- GraphVid overcomes limitations in complex scenes and occlusions
- It uses graph conditions to control video generation, enhancing interaction prec
- Users can flexibly control multiple objects through structured interactions
本页为 gitzw.com 基于公开来源的 AI 中文解读，非原文转载。