超逼真麦当劳生活方式肖像（两场景）
一个结构化的提示，包含两个独立的图像生成动作，使用 DALL-E（通过 Nano Banana Pro）来创建麦当劳场景中男性的超现实生活肖像。第一个场景是“温暖的一口”（微笑，拿着薯条），第二个场景是“沉思的凝视”（望着窗外）。这两个提示都强调使用上传的面部图像作为身份参考。
nano-banana-pro · 产品营销 · 人像 / 自拍 · 摄影 · 食品 / 饮料
中文提示词
图片 1: 温暖的一口 { "action": "dalle.text2im", "action_input": "{ \"prompt\": \"一张高质量、超逼真的照片，照片中的男子来自上传的图片（中等肤色，深色头发梳理整齐，胡须修剪得当）。他穿着一件干净的白色长袖毛衣，坐在一张明亮的麦当劳木桌旁。他对着镜头温暖地微笑，一只手拿着一个黄色的麦当劳杯子，另一只手拿着一根薯条。桌上放着一盒经典的红色麦当劳薯条。背景柔和模糊，带有金拱门标志和柔和的日光。\", \"size\": \"1024x1024\" }", } 图片 2: 若有所思的凝视 { "action": "dalle.text2im", "action_input": "{ \"prompt\": \"一张高质量、超逼真的生活照，照片中的男子来自上传的图片（中等肤色，飞机头，短胡须）。他穿着一件白色圆领毛衣，坐在麦当劳的靠窗座位上。他若有所思地望着窗外，带着一丝微笑。桌上放着一个黑色托盘，上面有一个装在盒子里的芝麻汉堡和薯条。温暖的自然光线，浅景深，餐厅内部模糊。\", \"size\": \"1024x1024\" }", Use my uploaded face image only as a face and identity reference. No other faces, models, datasets, or references are allowed. }
原始提示词
Image 1: The Warm Bite { "action": "dalle.text2im", "action_input": "{ "prompt": "A high-quality, ultra-realistic photo of the man from the uploaded image (medium skin tone, dark hair styled neatly, groomed beard). He is wearing a clean white long-sleeve sweater, sitting at a wooden table inside a brightly lit McDonald’s. He is smiling warmly at the camera, holding a yellow McDonald's cup in one hand and a single french fry in the other. A classic red box of McDonald's fries is on the table. Softly blurred background with the golden arches logo and soft daylight.", "size": "1024x1024" }", } Image 2: The Thoughtful Gaze { "action": "dalle.text2im", "action_input": "{ "prompt": "A high-quality, ultra-realistic lifestyle photo of the man from the uploaded image (medium skin tone, pompadour hairstyle, short beard). He is dressed in a white crewneck sweater, sitting at a window seat in a McDonald’s. He is looking thoughtfully out the window with a slight smile. On the table is a black tray with a sesame seed burger in a box and french fries. Warm natural lighting, shallow depth of field with a blurred restaurant interior.", "size": "1024x1024" }", Use my uploaded face image only as a face and identity reference. No other faces, models, datasets, or references are allowed. }