该提示词引导AI从图像中提取色彩、字体、构图、视觉效果与整体风格等信息,并以严格有效的JSON格式输出,字段可灵活增减。适合设计风格分析、AI绘画风格迁移与设计系统沉淀。
中文版提示词
请从输入图像中智能、全面地提取视觉风格信息,并以严格有效的JSON格式输出。字段数量不做限制,可根据图像特征灵活增减,但需保持结构清晰、语义明确、分类合理。建议结构(可动态调整、增删):colors(palette色板、dominant_colors主色、accents点缀色、tone_contrast明度/色温/对比)、typography(fonts字体、style_features字重/字宽/字型、hierarchy排版层级)、composition(layout布局、balance对称/非对称/中心构图、focal_points视觉焦点、spacing_and_rhythm留白/节奏/密度)、visual_effects(textures纹理、lighting光影、shadows阴影、filters滤镜、other_effects其他)、overall_style(design_language设计语言、emotional_tone感性气质、reference_genres类似风格类型)。要求:输出必须是纯JSON,不包含任何额外说明文字;可根据图像内容自由扩展或删减字段,但命名专业、语义明确;无法判断的字段使用空字符串、空数组或省略。
英文版提示词
Intelligently and comprehensively extract visual style information from the input image and output it in strictly valid JSON. The number of fields is not limited and can be flexibly added or removed based on the image, but the structure must be clear, semantically precise and well-categorized. Suggested structure (adjustable): colors (palette, dominant_colors, accents, tone_contrast), typography (fonts, style_features, hierarchy), composition (layout, balance, focal_points, spacing_and_rhythm), visual_effects (textures, lighting, shadows, filters, other_effects), overall_style (design_language, emotional_tone, reference_genres). Requirements: output must be pure JSON with no extra explanatory text; fields may be freely extended or removed, with professional naming and clear semantics; use empty strings, empty arrays, or omit fields that cannot be determined. 🛠️ **适用 AI 工具**:ChatGPT、Claude、Kimi、Gemini、通义千问、豆包

◯ 评论 0