Text-to-Image
People & Space
ChatGPT
Character Description
Prompt
【Task】
Generate a high-precision character sheet based on the reference image to lock the character ID. No new characters are allowed; all visuals must be based on the same character structure.
【Basic Settings】
Style: [Realistic 3D/Stylized 3D/Anime/Semi-realistic/IP Design]
Character Description: [Fill in your character description or upload a reference image]
Gender: [Male/Female/Neutral]
Age: [Value]
Body Type: [Slim/Standard/Muscular/Exaggerated Proportion]
Style Keywords: [High-end/Fashionable/Trendy/Tech-savvy/Emotional, etc.]
【Visual Structure】
- Aspect Ratio: 4:3 horizontal
- Background: Pure white/Off-white/Minimalist
- UI: Clean technical layout, no logo, no watermark
- Font: Clear and readable English labels
【Required Modules】
1. Top Information
- Name (can be auto-generated)
- Character Identity
- Age
- Personality Keywords (3-5)
- Core Theme (1 sentence)
2. Color System - 6-8 color blocks (no text)
3. Main Identity Display (largest area) Key: Lock character
- Front/3/4/Side/Back views
- Standard standing pose
- With proportion lines (height scale)
- No props
4. Silhouette
- Front silhouette
- Side silhouette
5. Expression System (8 images)
- Calm/Curious/Tense/Surprised/Scared/Sad/Determined/Relaxed
6. Micro-expressions (5 images)
- Eye tension/Smile/Mouth effort/Slight fear/Breath control
7. Head Structure
- Multiple angles (3/4/side/upward/downward)
8. Pose Variations
- Relaxed/Tense/Confident
9. Close-up Shot (1 image)
- Above chest
- Strong emotional expression
10. Clothing Details (4 images)
- Hairstyle/Material/Accessories/Shoes
11. Hand Gestures
- Relaxed/Tense/Pointing/Grabbing/Facial actions
【Consistency Requirements】
- All visuals must feature the exact same character (face/hairstyle/proportions/clothing)
- No style drift allowed
- Main display area must be the largest
【Quality Requirements】
- OG-level detail
- Realistic materials (skin/fabric/metal)
- Cinematic lighting and shadows
Translated from Chinese by AI.
Show the original prompt
【任务】
基于参考图生成一张高精度角色设定板(Character Sheet)锁定角色ID,不允许生成新角色,所有画面必须基于同一角色结构
【基础设定】
风格:[写实3D/风格化3D/动漫/半写实/IP设计]
角色描述:[填写你的角色描述或上传参考图]
性别:[男/女/中性]
年龄:[数值]
体型:[瘦/标准/健壮/夸张比例]
风格关键词:[高级感/时尚/潮流/科技感/情绪化等】
【画面结构】
- 画面比例:4:3横版
- 背景:纯白/米白/极简
- U1:干净技术排版,无logo,无水印
- 字体:清晰可读英文标签
【必须包含模块】
1. 顶部信息
- 名字(可自动生成)
- 角色身份
- 年龄
- 性格关键词(3-5个)
- 核心主题(1句)
2. 配色系统-6~8个色块(无文字)
3. 主身份展示(最大区域) 重点:锁定角色
- 正面/3/4/侧面/背面
- 标准站姿
- 带比例线(身高刻度)
- 无道具
4. 轮廓剪影
- 正面剪影
- 侧面剪影
5. 表情系统(8张)
- 平静/好奇/紧张/惊讶/害怕/悲伤/坚定/放松
6. 微表情(5张)
- 眼部紧张/微笑/嘴部用力/微恐惧/呼吸控制
7. 头部结构
- 多角度(3/4/侧面/仰视/俯视)
8. 姿态变化
- 放松/紧张/自信
9. 特写镜头(1张)
- 胸部以上
- 强情绪表达
10. 服装细节(4张)
- 发型/材质/配饰/鞋
11. 手部动作
- 放松/紧张/指向/抓握/面部动作
【一致性要求】
- 所有画面角色完全一致(脸/发型/比例/服装)
- 不允许风格漂移
- 主展示区域必须最大【质量要求】
- 0G级细节
- 材质真实(皮肤/布料/金属)
- 影视级光影





