迷你模型场景重建

A structured JSON prompt for GPT Image 2 to reconstruct any input image as a clean, collectible-scale miniature diorama.

{
 "reference_images": {
  "scene_reference": "ATTACHED_IMAGE",
  "usage_rule": "使用此图像作为场景布局、物体放置和构图的精确来源。不要重新诠释或替换场景。"
 },
 "concept": {
  "type": "微缩模型重建",
  "intent": "将附加的场景重建为一个干净的、可收藏规模的模型，同时保留原始的空间关系"
 },
 "environment": {
  "scale": "桌面微缩模型",
  "ground": {
   "type": "雕塑底座",
   "behavior": "自然地支撑所有物体，无需封装",
   "edges": "干净、简约、略微倒角"
  },
  "background": {
   "type": "无缝摄影棚背景",
   "color": "纯白色",
   "constraint": "除模型底座外无其他环境"
  }
 },
 "objects_and_elements": {
  "layout_rule": "所有物体必须保持与参照图片相同的相对位置",
  "primary_elements": "参照中可见的精确车辆、结构、道具和地形",
  "detail_handling": "简化但忠实的微缩比例",
  "characters": {
   "presence": "仅在参照中存在时",
   "style": "微缩人偶，无面部写实感"
  }
 },
 "composition": {
  "view": "等距或略微擡高的3/4视角",
  "framing": "整个模型可见，居中",
  "cropping": "无缩放，无电影式裁切"
 },
 "camera": {
  "lens_equivalent": "35mm-50mm",
  "perspective": "微缩真实感",
  "distortion": "无"
 },
 "lighting": {
  "type": "柔和的摄影棚灯光",
  "direction": "顶部并略微倾斜",
  "shadows": "仅物体下方有细微接触阴影",
  "avoid": ["戏剧性灯光", "聚光灯", "电影对比度"]
 },
 "materials_and_style": {
  "surface_finish": "光滑、哑光或轻微缎面",
  "material_quality": "高品质比例模型材料",
  "aesthetic": "干净、现代、高级的模型"
 },
 "quality_controls": {
  "image_clarity": "极度清晰",
  "grain": "无",
  "noise": "无",
  "motion_blur": "无"
 },
 "negative_prompts": [
  "玻璃罩", "展示盒", "博物馆展柜",
  "透明方块", "保护外壳",
  "玻璃反光", "标签或铭牌",
  "文本覆盖", "杂乱背景", "电影模糊"
 ]
}

机器翻译 — 实际生成作品的是原文提示词。

✨ 生成 收藏 复制提示词 分享 原始帖子 在 Hitboard 打开