原始文档
GPT Image 2 · en
A striking artistic portrait, with the subject being{argument name="subject" default="charming women"} (See attached image) Her facial features are precisely captured; she sits with a serene yet powerful posture. She sits cross-legged on a minimalist floor covered with decorative tiles, wearing a modern dress with a minimalist print.{argument name="top" default="blue sweatshirt"} , pairing{argument name="bottom" default="Loose jeans"} Wearing black Converse sneakers, her expression was serene and subtly confident, a gentle smile playing on her lips, her eyes relaxed, gazing softly at the diffused light. Behind her stood a massive abstract mural depicting her own face, bold and expressive in its use of charcoal gray, white, gold, and soft red. This mural was not a realistic copy, but a stylized and emotionally charged interpretation of her: deep-set eyes, fragmented textures, and broad brushstrokes hinting at movement, ambition, and inner complexity. The mural's scale dominated the background without obscuring the subject, creating a harmonious balance between reality and artistic representation. The lighting was cinematic and captivating, with soft reflections at the edges, the overhead glow outlining her silhouette. Tiny dust particles floated lightly in the air, capturing the shimmering light and enhancing the atmosphere. The scene blended a modern art studio with a dreamlike exhibition space: clean, spacious, and minimalist, yet brimming with emotional tension. Shot with hyper-realistic image quality using a 50mm portrait lens, featuring shallow depth of field, sharp details, and elegant color transitions, conveying themes of creativity, introspection, and personal transformation. Negative keywords: blurry, overexposed, cartoonish, anatomically distorted, superfluous limbs, poorly drawn hands, blurred faces, unrealistic colors, watermarks, text. Aspect ratio 9:16.使用方法
- 复制完整原始提示词，并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用，核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型，不代表该示例由这个模型生成。
- 核对名称、数字与画面文字；不同模型版本的结果可能不同。
