拍照前选择；同一剧情每次按固定节奏播放。双手靠近聚光，拉开形成彩虹；握拳碎裂，张开重新聚合。
SCRIPTED CONFIDENCE
FIXED SCENE DATA
94%
Coordinates describe the visible square in the mirrored photo. Area is its share of the image, not a confidence score.
The headline and 94 / 4 / 2 distribution are fixed script values. They remain identical across takes and are independent of the photograph and head position.
One mirrored frame is captured after the shutter is pressed.
A recent head-position box is preserved when available. No box is drawn when tracking is unavailable.
The camera stops before the animation. The authored scene output then appears.
Position detection runs on this device. No identity or sexual orientation is inferred. The photo stays in page memory and is cleared when the report closes or the session resets.
PROP SIMULATION