和这个灵感互动

点赞、评论与浏览数据保存在当前浏览器，无需登录。

本机记录

复制0

生成示例

产品营销 · 网页设计

面对面职业匹配合成图像生成器

一个复杂的 JSON 概念，描述基于参考面生成 4 面板合成图像。该合成图像包括一张大型“面部分析”肖像和三张较小的堆叠肖像，展示了人工智能根据感知的“面部氛围”和外观选择的三种不同职业的主题。

英文原文

Face-to-Career Matching Composite Image Generator

A complex JSON prompt for generating a 4-panel composite image based on a reference face. The composite includes a large 'face analysis' portrait and three smaller stacked portraits showing the subject in three distinct careers chosen by the AI based on the perceived 'facial vibe' and appearance.

第一次写图片提示词？不用理解英文变量，按照下面的中文步骤补充画面要求即可。

这个提示词需要上传参考图片

适合用途: 产品营销
建议模型: 通用模型
参考素材: 需要上传

开始填写这个模板

用这个模板开始创作

选择模式后，页面会自动整理成可复制的完整提示词。

1选择用途
2补充内容
3复制生成

为你整理的提示词

中文增强版

这条提示词没有可单独替换的变量，你仍然可以复制后直接修改。

变量替换结果

English

{
  "referenced_image_ids": ["<PASTE_INPUT_IMAGE_ID_HERE>"],
  "size": "1024x1536",
  "n": 1,
  "prompt": "Create a single composite image (a clean 4-panel collage) based on the provided reference face. The final output must contain: (A) one LARGE panel on the left (about 65% width, full height) showing a premium 'face analysis' portrait of the same person; and (B) three SMALL panels stacked vertically on the right (about 35% width) showing the SAME person in three different careers.\n\nCRITICAL CAREER-SELECTION RULE:\n- The three careers must be selected by YOU based on the person's facial vibe and overall appearance in the reference (e.g., perceived temperament, style, presence, confidence, softness/seriousness, elegance/athleticism, creative/analytical vibe). Do NOT pick random jobs.\n- Choose careers that feel plausibly aligned with the person’s look and presence (\"face-to-career matching\" concept).\n- The three careers must be DISTINCT and cover different archetypes (e.g., one corporate/strategic, one creative/media, one technical/field). Avoid repeating similar industries.\n- Keep the choices fun and flattering, but still believable.\n\nGLOBAL RULES (apply to all panels):\n- The person must look like the same individual in all panels (consistent facial features, hairline, eye shape/color, skin tone, proportions).\n- Realistic, high-end editorial photography style.\n- Do NOT add any real names, IDs, biometric scores, or identity claims. This is illustrative only.\n- Avoid heavy text. Minimal labels allowed only in the analysis panel.\n\nPANEL A (LEFT, LARGE): 'FACE ANALYSIS' PORTRAIT\n- Close-up, centered, symmetrical portrait (head and upper neck), studio lighting, sharp focus.\n- Overlay a semi-transparent facial analysis grid: symmetry axis, eye-line, nose base line, jaw curve arcs, proportional guides (golden ratio style).\n- Add subtle, minimal micro-labels in English such as: 'symmetry axis', 'eye line', 'jaw curve', 'proportion guide'. Keep labels tiny and tasteful.\n- Make it look like a premium aesthetic/cosmetic analysis diagram (modern UI overlay, not medical).\n\nPANELS B/C/D (RIGHT, THREE SMALL PANELS): CAREER PORTRAITS\nFor each of the three chosen careers, depict:\n1) A strongly job-accurate outfit (uniform/PPE/tools/accessories appropriate to the role).\n2) A clear environment that matches the job (office, studio, set, lab, site, outdoors, etc.).\n3) A light playful twist that fits the role (a charming prop, a witty moment, a fun-but-realistic action), without becoming cartoonish.\n- Keep the face visible and recognizable in each panel.\n- Use cinematic lighting and shallow depth-of-field; backgrounds should be rich but not cluttered.\n\nCOMPOSITION & QUALITY:\n- Output must be one single image containing all 4 panels with thin white borders between panels.\n- Cohesive color grading a"

来源原文

未修改

{
  "referenced_image_ids": ["<PASTE_INPUT_IMAGE_ID_HERE>"],
  "size": "1024x1536",
  "n": 1,
  "prompt": "Create a single composite image (a clean 4-panel collage) based on the provided reference face. The final output must contain: (A) one LARGE panel on the left (about 65% width, full height) showing a premium 'face analysis' portrait of the same person; and (B) three SMALL panels stacked vertically on the right (about 35% width) showing the SAME person in three different careers.\n\nCRITICAL CAREER-SELECTION RULE:\n- The three careers must be selected by YOU based on the person's facial vibe and overall appearance in the reference (e.g., perceived temperament, style, presence, confidence, softness/seriousness, elegance/athleticism, creative/analytical vibe). Do NOT pick random jobs.\n- Choose careers that feel plausibly aligned with the person’s look and presence (\"face-to-career matching\" concept).\n- The three careers must be DISTINCT and cover different archetypes (e.g., one corporate/strategic, one creative/media, one technical/field). Avoid repeating similar industries.\n- Keep the choices fun and flattering, but still believable.\n\nGLOBAL RULES (apply to all panels):\n- The person must look like the same individual in all panels (consistent facial features, hairline, eye shape/color, skin tone, proportions).\n- Realistic, high-end editorial photography style.\n- Do NOT add any real names, IDs, biometric scores, or identity claims. This is illustrative only.\n- Avoid heavy text. Minimal labels allowed only in the analysis panel.\n\nPANEL A (LEFT, LARGE): 'FACE ANALYSIS' PORTRAIT\n- Close-up, centered, symmetrical portrait (head and upper neck), studio lighting, sharp focus.\n- Overlay a semi-transparent facial analysis grid: symmetry axis, eye-line, nose base line, jaw curve arcs, proportional guides (golden ratio style).\n- Add subtle, minimal micro-labels in English such as: 'symmetry axis', 'eye line', 'jaw curve', 'proportion guide'. Keep labels tiny and tasteful.\n- Make it look like a premium aesthetic/cosmetic analysis diagram (modern UI overlay, not medical).\n\nPANELS B/C/D (RIGHT, THREE SMALL PANELS): CAREER PORTRAITS\nFor each of the three chosen careers, depict:\n1) A strongly job-accurate outfit (uniform/PPE/tools/accessories appropriate to the role).\n2) A clear environment that matches the job (office, studio, set, lab, site, outdoors, etc.).\n3) A light playful twist that fits the role (a charming prop, a witty moment, a fun-but-realistic action), without becoming cartoonish.\n- Keep the face visible and recognizable in each panel.\n- Use cinematic lighting and shallow depth-of-field; backgrounds should be rich but not cluttered.\n\nCOMPOSITION & QUALITY:\n- Output must be one single image containing all 4 panels with thin white borders between panels.\n- Cohesive color grading a"

图片与提示词已收录至画引媒体库 · 如有版权或下架需求，请联系我们。