Other: runway_act_one API
Act-One is an advanced performance-driven video generation model developed by Runway. It revolutionizes traditional performance capture by eliminating the need for specialized motion capture suits or complex rigging. Using just a standard driving video of a human performer and a single character reference image, Act-One transposes eye movements, facial expressions, micro-muscle twitches, and lip-syncing onto the target character. The model excels at capturing intense emotional performances while maintaining high character consistency, allowing creators to control cinematic or animated AI characters through direct acting.
- Input: image, video
- Output: video
- Released: 2024-10-21