OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image

东京夜色中的回眸抓拍风肖像

Candid Portrait of a Glance Back in Tokyo at Night

一位明确年满25岁的成年女性在东京夜间瓷砖墙旁自然迈步,回头看向镜头,一只手抬近镜头作轻微遮挡,带着羞涩克制的微笑。画面采用经本人知情同意的抓拍风演绎,保留半侧身、略偏的构图与手发局部动态拖影;手机闪光灯打亮人物,街道背景深暗,白色瓷砖墙拖成柔和光带,呈现真实而不过度失焦的夜间手机照片质感。

东京夜色中的回眸抓拍风肖像 — AI image creation
东京夜色中的回眸抓拍风肖像

提示词 / Prompt

中文提示词

创作一张竖幅9:16夜间手机街头照片,场景为东京一处安静的人行道,旁边是墙面整洁的白色瓷砖建筑。画面中是一位明确年满25岁的成年女性,使用配置人物参考ref-10a-v1。演绎一段经本人知情同意的自然抓拍风瞬间:她知道镜头存在,行走时自然回头。捕捉她迈步中的半侧身回眸,一只手抬起并部分靠近镜头,构图略偏,手部和头发带有受控的动态模糊;她神情羞涩克制、略带笑意,不表现惊恐或不适。手机直闪照亮面部与抬起的手,周边环境沉入深暗。白色瓷砖墙向后延伸,化作柔和的拖影光带。营造手持手机快照质感、可见颗粒、直闪和自然不均匀曝光,构图虽不完美但仍清晰易辨;保持面部可辨,动态模糊仅落在手部与飘动的发丝上,不要涂抹五官。整体自然随意、未经精修,但不暗示偷拍或侵扰。不增加其他人物、可读文字、标志或水印。

English prompt

Create a realistic vertical 9:16 nighttime smartphone snapshot on a quiet Tokyo nighttime sidewalk beside a building with clean white ceramic tile walls. Show one clearly adult woman, age 25 or older, using configured identity reference ref-10a-v1. Stage a consensual candid-style moment: she is aware of the camera and naturally turns her head back while walking. Capture her mid-step in a half-profile glance back, with one raised hand partly near the lens, slight off-center framing and controlled blur on the hand and hair; her expression is shy, restrained and lightly amused, never frightened or distressed. A direct phone flash lights her face and raised hand while the surroundings fall into deep darkness. White ceramic tiles recede behind her as softly streaked light bands. Apply a handheld smartphone-snapshot look, visible grain, direct flash and natural uneven exposure, with imperfect but legible framing; keep the face recognizable and reserve motion blur for the hand and loose hair, not smeared facial features. The result should feel spontaneous, casual and unpolished without implying covert photography or intrusion. No extra people, readable text, logos or watermark.

查看 API 模型

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

  • 质感拉满的高端商业美女广告

    以已核验的当前固定女性肖像保持面部身份,制作可替换产品与品牌氛围的高端商业广告;默认产品为咖啡、氛围为奢华,不依赖特定品牌或产品图。 Create a premium commercial ad using the verified configured woman’s portrait for face identity, with replaceable product and brand mood; default to coffee and luxury without requiring a specific brand or product image.

  • 高级极简人像海报设计

    以当前固定人像作为本次原始人物照片,制作3:4竖版高级极简肖像海报;保留人物可见身份、发型、发色、服装和姿势,并用可替换的三行英文默认文案。 Use the configured portrait as this preview’s original photo for a 3:4 minimalist poster, preserving visible identity and appearance with three replaceable English copy lines.

  • 超写实随性短发少女特写

    以当前固定人像中的人物面部身份为参考,创作暖光、随性自拍感的超写实短发特写;短波波头是来源明确要求的造型转换。 Create a warm, casual selfie-like hyper-realistic short-bob portrait using the configured portrait for face identity; the bob is the source-requested hairstyle transformation.

  • 家庭家务 4x4 灰度故事板

    制作包含16个连续家务步骤的4×4灰度3D教学故事板,使用已核验的人像参考统一角色面部身份;不声称参考图中未显示的全身比例已被精确匹配。 Create a monochrome 4×4 3D instructional storyboard of sixteen household chores, using the verified portrait reference to keep the character’s face and visible identity consistent without claiming exact unseen full-body proportions

  • 女性美妆分析的高清指南

    以已配置并核验的人像参考为主体,制作9:16女性美妆分析杂志风信息图;保留人物真实面部身份与自然肤质,按外观估计而非诊断皮肤类型。 Create a 9:16 editorial beauty-analysis infographic using the verified configured portrait as its subject; preserve the person’s facial identity and natural skin, and present skin type only as a visual estimate, not a diagnosis.

  • 高级商务证件照头像海报设计

    使用当前配置的人像参考制作正式商务头像海报,保留人物身份,并以可替换的姓名、职位和机构资料呈现英文信息栏。 Create a formal business headshot poster from the configured portrait reference, preserving the person’s identity and displaying replaceable name, position and institution details in an English information bar.