PROMPT RADAR / PROMPT平面设计

个人资料 / 头像 - Naive Illustration Portrait Transformation

A comprehensive prompt designed to transform a reference photo into a minimalist, contemporary naive art illustration while strictly preserving the subject's identity and composition.

@codewithhajra 0 热度 0 浏览
个人资料 / 头像 - Naive Illustration Portrait TransformationORIGINAL VISUAL / 来源见文末
编辑先看

它为什么值得进入灵感库

先别急着照抄整段 Prompt:真正有用的是主体、空间和镜头之间的约束关系。 先规定信息层级和文字安全区,再谈视觉风格;无法落版的漂亮画面不是完整设计。

完整视觉 Prompt

Transform the uploaded image into a premium {argument name="illustration style" default="contemporary naïve illustration"} while preserving the exact facial identity, facial structure, hairstyle, expression, body proportions, pose, clothing silhouette, composition, and overall scene from the reference image.\nFORMAT LOCK\nVertical 3:4 composition. Preserve the original framing, camera angle, perspective, subject placement, and overall composition. Clean centered layout with a premium poster aesthetic.\nIDENTITY LOCK\nPreserve the exact facial features, hairstyle, expression, body proportions, clothing silhouette, and recognizable identity. Simplify the illustration without changing the person’s appearance.\nILLUSTRATION STYLE\nCreate a contemporary authorial naïve illustration with a minimalist hand-drawn aesthetic. Use playful proportions, slightly irregular hand-drawn outlines, expressive decorative linework, and intentionally imperfect craftsmanship. Simplify forms while keeping the subject immediately recognizable.\nCOLOR SYSTEM\nUse flat high-contrast color blocks with a refined contemporary palette. No gradients. No shadows. No realistic lighting. No 3D effects. Clean graphic shapes with bold visual balance.\nSTYLE\nPremium designer poster. Editorial illustration. Modern naïve art. Handmade graphic illustration. Museum-quality print design. Minimalist composition. Contemporary art-book aesthetic. Expressive handcrafted character. Ultra-clean premium artwork.\nNEGATIVE PROMPT\nAvoid photorealism, gradients, shadows, 3D effects, CGI appearance, digital painting, excessive detail, messy linework, blurry edges, distorted anatomy, extra limbs, text, logos, watermarks, or low-quality rendering.
查看原作者内容

不要一次改完所有变量

用 GPT Image 2 复刻时,先固定主体和构图,再逐次替换光线、色调或材质中的一个变量,这样更容易判断哪项描述真正起作用。

  1. 主体:A comprehensive prompt designed to transform a reference photo into a minimalist, contemporary naive art illustration while strictly preserving the subject's identity and composition.
  2. 视觉方向:社区实测摄影提示词;按原提示词的光线描述执行
  3. 镜头与构图:按原提示词的镜头与景别执行;按原提示词的构图与画幅执行
  4. 交付约束:auto,遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。
负向约束

遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。

模型建议

GPT Image 2 / Grok Imagine / Midjourney / Flux。先用原始 Prompt 建立基准结果,再根据模型对自然语言、镜头运动或文字排版的能力做局部改写。

来源与编辑原则

保留作者,也保留判断。

本页收录公开案例并提供中文索引、结构化拆解和复刻建议,不冒充原作者作品,也不替代原帖上下文。

作者:H A J R A · 收录于 2026年8月11日

访问原始来源

沿着这条线继续看