摄影写实 / PROMPT
樱花咖啡户外人像
严格控制人物姿态、服装、镜头和粉色调,生成明亮自然的户外生活方式写真。
适用模型GPT Image 2
已实测灵感来源于 awesome-gpt-image-2 案例 377;当前提示词与效果图由 Fastgoing 使用 GPT Image 2 实测重新制作。
完整提示词COPY / USE
Edit the provided image while preserving the same face identity, shape, and facial features without altering age, ethnicity, or structure. Maintain a calm, relaxed expression with the subject not looking at the camera.
Subject: young woman (18–23) with soft feminine beauty, smooth glowing skin, natural texture.
Pose (strict): seated on a wooden chair, body angled sideways (45–70°), legs crossed naturally, upper body slightly leaning forward, head turned away from camera, gaze to the side. She holds a drink cup with a straw using both hands.
Framing: full-body vertical (9:16), head to shoes visible, centered but slightly offset for natural composition.
Outfit: light pink varsity jacket with white stripes, soft pink inner top, modest knee-length pleated skirt, white sneakers.
Accessories: beige newsboy cap, sunglasses (on face or cap), pink shoulder bag, small earrings.
Hair: neat low bun with soft loose strands.
Environment: outdoor flower shop street scene with pastel flowers (pink, soft tones), decorative plants, floral storefront.
Lighting: bright natural daylight, soft glossy skin highlights, balanced exposure, soft shadows.
Camera: low/frog angle (slightly below, looking up), 50mm lens, shallow depth of field.
Color grading: warm pastel palette (pink, cream), clean bright lifestyle aesthetic.
Quality: ultra-photorealistic, 8K detail, DSLR realism, natural skin texture.
Negative: front-facing, eye contact, close-up, cropped body, mini/short skirt, indoor scene, dark lighting, anime, cartoon, CGI, plastic skin, distorted anatomy.
--ar 9:16 --style raw --quality high复制后,可参考下方“建议替换”项目改成你的主题;带【方括号】的内容请优先替换。
PROVENANCE / 来源与验证
它从哪里来,是否完成实测?
灵感来源于 awesome-gpt-image-2 案例 377;当前提示词与效果图由 Fastgoing 使用 GPT Image 2 实测重新制作。
GPT Image 2 · 2026-08-17 实测HOW TO USE
不用学提示词工程,也能改成自己的。
建议替换
01【参考人物】02【服装配色】03【手持道具】04【户外场景】05【镜头与画幅】使用小贴士
先只替换变量跑一次,观察最不符合预期的部分;第二轮只补充那一项约束,通常比一次塞进很多要求更稳定。
