GPT-image-2 热门玩法实战（二）：AI 面相分析个人色彩诊断

GPT-image-2 热门玩法实战（二）：AI 面相分析 & 个人色彩诊断 — 上传自拍秒出专业报告

系列导读：本系列通过 Crazyrouter API 实战演示 GPT-image-2 的各种病毒式玩法。上一篇我们玩了 AI 看手相，这一篇来搞两个同样火爆的玩法：AI 面相分析和个人色彩季节诊断。

🔥 这两个玩法为什么火？

AI 面相分析：上传一张正面照，GPT-image-2 会生成一张标注了面部特征的分析图 — 额头、眉形、眼型、鼻型、嘴唇、下巴轮廓，并给出性格和运势解读。在亚洲社交媒体上尤其火爆。

个人色彩诊断（Color Analysis）：这是 TikTok 和小红书上的超级热门话题。传统做法需要花几百块找专业色彩顾问，现在 GPT-image-2 一张图就能告诉你属于哪个色彩季节（春/夏/秋/冬），以及最适合你的穿搭配色。

🎨 效果展示

下面是通过 Crazyrouter API 调用 GPT-image-2 实际生成的效果图：

面相分析效果图

个人色彩诊断效果图

📝 面相分析 — 完整代码

Python 版本

fromopenaiimportOpenAI client=OpenAI(api_key="your-crazyrouter-api-key",base_url="https://crazyrouter.com/v1")# 面相分析 promptface_reading_prompt=""" Create a professional Chinese face reading (面相分析) infographic. Layout: A front-facing portrait in the center with annotation lines pointing to key facial features. Analyze and label these features: - 额头 (Forehead): shape and height → career fortune - 眉形 (Eyebrows): thickness and arch → personality traits - 眼型 (Eyes): shape and spacing → wisdom and relationships - 鼻型 (Nose): bridge and tip → wealth fortune - 嘴唇 (Lips): shape and fullness → communication style - 下巴 (Chin): shape and prominence → later life fortune - 脸型 (Face shape): overall → life overview Style: dark navy background, gold and white text, elegant Chinese aesthetic. Include a summary section at bottom with: 事业运 (Career), 财运 (Wealth), 感情运 (Love), 健康运 (Health). Each rated with 1-5 stars. Title: 面相分析报告 | AI Face Reading Report """response=client.images.generate(model="gpt-image-2",prompt=face_reading_prompt,size="1024x1024",n=1)print(f"面相分析图：{response.data[0].url}")

curl 版本

curl-XPOST https://crazyrouter.com/v1/images/generations\-H"Content-Type: application/json"\-H"Authorization: Bearer your-crazyrouter-api-key"\-d'{ "model": "gpt-image-2", "prompt": "Create a professional Chinese face reading infographic. Dark navy background, gold accents. Center portrait with labeled facial features: forehead (career), eyebrows (personality), eyes (wisdom), nose (wealth), lips (communication), chin (later life). Bottom summary with star ratings for Career, Wealth, Love, Health. Title: 面相分析报告 AI Face Reading Report.", "size": "1024x1024", "n": 1 }'

📝 个人色彩诊断 — 完整代码

Python 版本

fromopenaiimportOpenAI client=OpenAI(api_key="your-crazyrouter-api-key",base_url="https://crazyrouter.com/v1")color_analysis_prompt=""" Create a professional Personal Color Season Analysis infographic. Layout: 4 quadrants, one for each color season: - 🌸 Spring (春): warm, bright, clear — coral, peach, warm yellow, grass green, turquoise, cream - ☀️ Summer (夏): cool, muted, soft — lavender, dusty rose, powder blue, sage green, mauve, silver - 🍂 Autumn (秋): warm, deep, rich — burnt orange, olive green, mustard, burgundy, chocolate, rust - ❄️ Winter (冬): cool, vivid, high-contrast — pure white, black, royal blue, emerald, hot pink, icy silver Each quadrant shows 6 color swatches with hex codes. Center: a stylized female silhouette. Highlight the Autumn quadrant with a golden border and "YOUR BEST MATCH ✓" label. Bottom section: - Best colors to wear - Colors to avoid - Makeup palette recommendations - Jewelry metal recommendation (gold vs silver) Style: clean, modern, fashion magazine editorial. Title: YOUR COLOR SEASON ANALYSIS | 个人色彩诊断 """response=client.images.generate(model="gpt-image-2",prompt=color_analysis_prompt,size="1024x1024",n=1)print(f"色彩诊断图：{response.data[0].url}")

💡 Prompt 优化技巧

面相分析

指定文化背景：中国面相学和西方面部分析侧重点不同，prompt 中明确 “Chinese face reading” 或 “physiognomy”
星级评分：加入量化评分让结果更直观
双语标注：中英双语适合国内外用户

色彩诊断

明确四季体系：Spring/Summer/Autumn/Winter 是标准的四季色彩理论
给出具体色号：要求 hex code 让结果更实用
加入穿搭建议：不只是色卡，还有实际应用场景

🎯 进阶玩法：结合上传照片

GPT-image-2 支持图片输入。如果你有用户的自拍照，可以结合照片做更个性化的分析：

importbase64# 读取用户照片withopen("user_selfie.jpg","rb")asf:image_b64=base64.b64encode(f.read()).decode()# 结合照片的 promptprompt=f""" Based on this person's facial features, create a personalized face reading analysis. Annotate their actual facial features with lines and labels. Include personality traits and fortune predictions based on Chinese physiognomy. Dark elegant design with gold accents. """# 注意：图片输入需要使用 chat completions API 配合 image_url# 纯 images/generations 端点目前主要支持文本输入