微缩模型场景重构

一个用于 GPT Image 2 的结构化 JSON 提示词，旨在将任何输入图像重构为干净、可收藏级别的微缩模型场景。

提示词

{ "reference_images": { "scene_reference": "ATTACHED_IMAGE", "usage_rule": "使用此图像作为场景布局、物体摆放和构图的精确来源。请勿重新诠释或替换场景。" }, "concept": { "type": "微缩模型场景重构", "intent": "将所附场景重构为干净、可收藏级别的微缩模型，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": "自然支撑所有物体，无需封闭", "edges": "干净、简约、略带斜角" }, "background": { "type": "无缝摄影背景", "color": "纯白", "constraint": "底座之外无任何环境" } }, "objects_and_elements": { "layout_rule": "所有物体必须保持与参考图像中相同的相对位置", "primary_elements": "参考图像中可见的精确车辆、结构、道具和地形", "detail_handling": "简化但忠实于微缩比例", "characters": { "presence": "仅在参考图像中存在时包含", "style": "微缩人偶，无需面部写实" } }, "composition": { "view": "等轴测或略微抬高的 3/4 视角", "framing": "完整显示整个微缩模型，居中", "cropping": "无缩放，无电影感裁剪" }, "camera": { "lens_equivalent": "35mm-50mm", "perspective": "微缩写实感", "distortion": "无" }, "lighting": { "type": "柔和摄影棚灯光", "direction": "头顶上方并略带角度", "shadows": "仅物体下方有细微的接触阴影", "avoid": ["戏剧性灯光", "聚光灯", "电影感对比度"] }, "materials_and_style": { "surface_finish": "光滑、哑光或轻微缎面", "material_quality": "高质量比例模型材质", "aesthetic": "干净、现代、高级微缩模型" }, "quality_controls": { "image_clarity": "超清晰", "grain": "无", "noise": "无", "motion_blur": "无" }, "negative_prompts": [ "玻璃罩", "展示盒", "博物馆封闭罩", "透明立方体", "保护壳", "玻璃反光", "标签或铭牌", "文字叠加", "杂乱背景", "电影感模糊" ] }