前の画像を次の画像のように単体の被写体に
Want a different result? Change just a few parts
AI finds the swappable parts (subject/style/scene) — fill in what you want, keep the rest