基于深度视频的 K-pop 舞蹈动作映射

“Use Image 1 for the woman's appearance and clothes and Image 2 for the man's, replacing the setting with a neon-lit Seoul street at night. Follow the supplied depth-reference video for the full choreography: movements, positions, body posture, interactions, entrance order and rhythm should remain matched rather than reinterpreted. Begin with the woman speaking a brief reminder, starting in Mandarin and ending in natural Korean, with each language synchronized to her mouth. Continue the reference sequence as a single vertical take with a polished K-pop music-video feel. Preserve both identities, clothing and limb continuity throughout; leave out subtitles, other text and watermarks.”