実在の誰かに見えない人物写真風の画像を出すプロンプトです。 [何をしているか]を書き換えて画像生成モデルに貼ります。資料に人物写真が要るときに。有名人名を入れないだけでは似た顔が出ます。顔が主役にならない構図にするのが確実で、後ろ姿や横顔に寄せます。
- Midjourneyバージョン指定なし
a candid photograph of a person [何をしているか], seen from behind at a slight angle, face not visible, natural indoor lighting from a window on the left, shallow depth of field, background softly out of focus, neutral everyday clothing, no logos on clothing, documentary style, unposed, natural posture, 35mm lens, eye level, slightly off-center composition --no face, no facial features, no direct eye contact, no celebrity likeness, no recognizable person, no brand logo, no text, no watermark, no studio lighting, no exaggerated expression 【使い方】 1行目の [何をしているか] を具体的に書く(例:typing on a laptop at a cafe table)。 行為が具体的なほど、顔に頼らずに状況が伝わる。 【顔が要るとき】 seen from behind を次のどれかに変える。特定の誰かに寄るのを避けやすい順。 - seen from the side, profile view, eyes looking away from camera - seen from a distance, small in frame, surrounded by environment - hands and workspace only, person cropped out above the shoulders 【注意】 実在の人物名・キャラクター名をプロンプトに入れない。 似ている人が出た場合は、その画像を使わずに作り直す。