A reproducible vertical dance outfit-change generation package built from one reference dance video, one adult female model image, six outfit images and one scene image. The system first parses the action time nodes, then generates six single-person outfit-change images, and finally composites in a fixed order a video of the same model dancing while changing outfits in the specified scene. What you ultimately get from this Skill: 1. parse the reference video and generate action and time nodes; 2. run "outfit image plus model image" one by one to get outfit-change images 1-6; 3. confirm the character identity, outfit order and frame safety of the six process images; 4. enter the final video node with the reference video, scene image, outfit-change images 1-6 and time nodes; 5. use fixed parameters to generate a 9:16, 13-second, 720p vertical video.
This course comes with a materials pack. Members receive the access link by email after requesting the course.