OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image
现代建筑广场超现实多重比例时尚大片
Surreal Multi-Scale Fashion Editorial in a Modern Architectural Plaza
现代建筑广场中的超现实多重比例时尚广告，同一成年女性以近景、巨型建筑后身影和清晰广告肖像三种尺度出现。
提示词 / Prompt
中文提示词
以用户上传的参考图作为主要可见人物的面部身份依据。在不与身份参考冲突的情况下，保留来源的场景、造型、服装、姿势、构图、光线和视觉风格。 创作一幅明亮现代建筑广场中的超现实高端时尚广告照片，采用4:5竖幅构图。the same clearly adult woman in all three appearances恰好出现三次、分别处于明显不同的尺度；三处都保持相同可辨认的面孔、发型与成年形象。前景近处是一位正常比例人物，穿白色短袖衬衫和深色条纹领带，一只手靠近下巴；面部保持可读，只允许轻微景深柔化。大型白色石材博物馆体量后方露出同一女性的巨型版本，只见上半张脸、头部和一只手，面向镜头并指向广告牌。博物馆立面上设置一块大型竖版时尚广告牌，清晰呈现她穿同一套服装的锐利肖像。 背景采用淡蓝天空、白色几何建筑与玻璃窗，配真实日光和柔和阴影。通过一致透视、接触阴影和前景、建筑、广告牌之间的清晰层次，让超现实尺度关系一眼可辨。广告牌不放文案。保留自然肤质与发丝细节；避免身份漂移、面部扭曲、额外人物、标志、水印、随机文字或建筑变形。
English prompt
Use the uploaded reference image as the facial identity reference for the main visible person. Keep the source scene, styling, clothing, pose, composition, lighting and visual style unless they conflict with the identity reference. Create a surreal high-fashion campaign photograph in a bright modern architectural plaza, using 4:5 portrait framing. Show the same clearly adult woman in all three appearances exactly three times at distinct scales, with the same recognizable face, hairstyle and adult appearance in every instance. In the foreground, include a normal-size version near the camera in a white short-sleeve shirt and dark striped tie, with one hand near the chin; keep her face readable with only gentle depth-of-field softness. Behind a large white-stone museum volume, reveal an oversized version of the same woman—upper face, head and one hand visible—looking toward camera and pointing to the billboard. On the museum facade, place a large vertical fashion billboard with a sharply focused portrait of her in the same outfit. Use pale blue sky, white geometric architecture, glass windows, realistic sunlight and soft shadows. Make the impossible scale relationships immediately legible through consistent perspective, contact shadows and clean separation of foreground, building and billboard. Keep the billboard free of copy. Preserve natural skin and hair detail; no identity drift, face distortion, extra people, logos, watermark, random text or warped architecture.
探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作
相关作品
- 90年代冷色调仙侠风翡翠绿汉服人像
Ethereal 9:16 full-body portrait of a clearly adult Chinese woman in jade-green xianxia-inspired hanfu amid layered gauze curtains, silver ornaments, and cold teal moonlight with nostalgic 1990s fantasy-drama diffusion.
- 超精细3D漫画球星收藏级海报
Reusable 4:5 sports-memorabilia poster template for a named football player, combining a stylized likeness, current kit, chronological miniature career display, trophies, milestone props, and engraved nickname plaque.
- 复古赛车海报设计提示词
Reusable 4:5 vertical retro racing poster template with a centered front-facing car hero view, integrated geometric landscape graphics, fictional brand mark, slogan, and distressed vintage print finish.
- Q版角色聊天表情包生成指令
Reusable template for a consistent chibi chat-sticker character showing one clear emotion and action at a time, with a reserved Chinese-copy area and white or transparent background.
- 认识蜜蜂儿童科普百科海报
Colorful classroom infographic poster centered on a friendly fuzzy honey bee, with modular anatomy, life-cycle, superpower, and conservation sections in a soft cream honeycomb design.
- 时尚色块拼贴人像摄影
Premium fashion-campaign collage of five adult European models in separate rounded geometric panels, using royal blue, crimson, emerald, mustard, and white luxury color fields with clean studio flash.