用于生成写实女性人物肖像图，融合三张参考图的面部特征（首图定脸型40%、二三图定五官神态60%），强调冷艳清冷气质与中亚/东欧异域风情，含负面提示词规避变形，适合角色形象定制。

中文版提示词

生成一张写实女性肖像图。基础设定：年龄约22-25岁，身高168cm，体型纤细修长（50kg），冷艳可爱、清冷气质，带微妙中亚或东欧异域风情，皮肤白皙，五官立体精致，眼神柔和但疏离，唇形优美。

面部基因融合要求：融合三张参考图的面部特征，以第一张图为脸型基础（权重40%），融合第二、三张图的五官细节与神态（总权重60%），生成一张全新的、有辨识度的脸。

画面控制：纯白背景，专业影棚灯光，清晰对焦，8K UHD，胶片颗粒质感。

负面提示词（必须）：(worst quality, low quality, normal quality:1.4), deformed, mutated, disfigured, blurry, watermark, text, signature, extra limbs, bad anatomy, ugly, cartoon, 3d, render, doll。

正向前缀（如使用Lora模型）：<lora:你的风格Lora模型（如有）:0.8> 1girl, (masterpiece, best quality, ultra-detailed, realistic photo), a beautiful young woman, face portrait, looking at viewer。

英文版提示词

Generate a photorealistic female portrait. Base settings: about 22-25 years old, 168 cm tall, slender (50 kg), cool and lovely with a detached aura and a subtle Central Asian or Eastern European exotic flavor; fair skin, refined three-dimensional features, gentle but aloof eyes, and elegant lips.

Face-fusion requirement: blend facial features from three reference images — use the first image as the face-shape base (weight 40%) and fuse the facial-feature details and expression of the second and third images (total weight 60%) to create a brand-new, recognizable face.

Image control: pure white background, professional studio lighting, sharp focus, 8K UHD, film grain.

Negative prompt (mandatory): (worst quality, low quality, normal quality:1.4), deformed, mutated, disfigured, blurry, watermark, text, signature, extra limbs, bad anatomy, ugly, cartoon, 3d, render, doll.

Positive prefix (if using a LoRA): <lora:your style LoRA model (if any):0.8> 1girl, (masterpiece, best quality, ultra-detailed, realistic photo), a beautiful young woman, face portrait, looking at viewer.

🛠️ **适用 AI 工具**：Midjourney、Stable Diffusion、DALL·E、即梦、豆包、Nano Banana