在生成之前先上传一张参考图像。
分享链接 · https://openimages.ajiang.me/p/2103157080370684198/


composition · centeredcomposition · full-bodyhuman_attributes · curly-hairhuman_attributes · glassessubject_type · charactersubject_type · fashion-person
提示词(原文,逐字保留)
UPLOAD A REFERENCE IMAGE BEFORE GENERATING.
MAIN PROMPT
Using the uploaded photo as the primary reference, transform ONLY the main subject into a charming oversized-head hand-drawn cartoon caricature. The head must be dramatically exaggerated, approximately 45–60% of the character’s visual height, with a compact, much smaller body. Preserve the subject’s recognizable identity and defining characteristics while keeping the original photograph and environment fully recognizable.
FORMAT LOCK
Vertical 3:4. Preserve the original framing, camera angle, perspective, depth, subject placement, pose, orientation, and composition. Keep the photographic environment unchanged.
STYLE TRANSFER
Create a naïve children’s-book caricature with a deliberately huge, subject-specific head, enlarged forehead and cheeks, tiny simplified body, short limbs, small hands and feet, and a cute humorous proportion. Adapt the head shape, silhouette, hairstyle, hairline, facial hair, ears, beak, hat, helmet, or other defining features specifically to the original subject. Never use a generic circular head.
DETAIL & COLOR
Preserve facial structure, expression, skin tone, hairstyle, facial hair, clothing, accessories, patterns, logos, shoes, jewelry, and distinctive details. Use simple dot/oval eyes, tiny simplified nose, curved mouth, subtle cheeks when appropriate, soft natural colors, visible colored-pencil strokes, rough hatching, cross-hatching, uneven fills, overlapping strokes, imperfect edges, and subtle paper texture.
COMPOSITION LOCK
Transform ONLY the main subject. Keep the entire background photorealistic, including architecture, landscape, furniture, vehicles, scenery, lighting, depth, perspective, and atmosphere. Integrate the illustrated character naturally into the original scene with matching position, perspective, lighting direction, and subtle grounding shadows.
STYLE
Loose hand-drawn ink and colored-pencil caricature, charming, whimsical, naïve, expressive, handmade, slightly imperfect. Thin-to-medium organic black outlines, rough pencil texture, spontaneous sketch marks, simplified clothing, exaggerated oversized head, tiny body. The contrast between the realistic photograph and illustrated subject should be intentional and visually striking.
NEGATIVE PROMPT
Normal head proportions, small head, generic head shape, generic cartoon character, identical head shape across subjects, photorealistic subject, cartoonized background, fully illustrated scene, anime, vector art, 3D CGI, Pixar, Disney, glossy digital rendering, plastic texture, smooth gradients, polished graphics, altered identity, changed hairstyle, changed clothing, changed pose, distorted facial features, excessive facial detail, oversized body, realistic anatomy, unnecessary background changes.
ASPECT RATIO: 3:4
中文译文
在生成之前先上传一张参考图像。
主提示
以上传的照片为主要参考,仅将主体人物转变为一个可爱的、手绘的大头卡通漫画形象。头部必须被戏剧性地夸张,约占角色视觉高度的 45–60%,身体紧凑且明显更小。在保留原始照片和环境完全可辨识的同时,保留主体的可识别身份和标志性特征。
格式锁定
竖向 3:4。保留原始构图、相机角度、透视、景深、主体位置、姿势、朝向和组成。保持摄影环境不变。
风格迁移
创作一个天真童趣的童书漫画形象,带有刻意放大的、主体专属的头部、扩大的额头和脸颊、极小而简化的身体、短小的四肢、小的手脚,以及可爱幽默的比例。根据原始主体专门调整头部形状、轮廓、发型、发际线、面部毛发、耳朵、喙、帽子、头盔或其他标志性特征。绝不要使用通用的圆形头部。
细节与色彩
保留面部结构、表情、肤色、发型、面部毛发、服装、配饰、图案、标志、鞋子、珠宝和独特的细节。使用简单的圆点/椭圆眼睛、极小而简化的鼻子、弯曲的嘴巴、在合适时的微妙脸颊、柔和的自然色调、可见的彩色铅笔笔触、粗糙的影线、交叉影线、不均匀的填充、重叠的笔触、不完美的边缘以及微妙的纸张纹理。
构图锁定
仅转变主体人物。保持整个背景的照片般写实,包括建筑、风景、家具、车辆、场景、光照、景深、透视和氛围。将插画角色自然地融入原始场景中,匹配位置、透视、光照方向和微妙的接地阴影。
风格
松散的手绘墨水和彩色铅笔漫画,可爱、奇思妙想、天真、富有表现力、手工制作、略带不完美。细至中等的有机黑色轮廓、粗糙的铅笔质感、自发的速写痕迹、简化的服装、夸张的大头、极小的身体。写实照片与插画主体之间的对比应当是刻意的且视觉上引人注目。
负面提示
正常的头部比例、小头部、通用头部形状、通用卡通角色、跨主体相同的头部形状、照片般写实的主体、卡通化的背景、完全插画的场景、动画、矢量艺术、3D CGI、皮克斯、迪士尼、光泽数字渲染、塑料质感、平滑渐变、精致图形、被改变的身份、被改变的发型、被改变的服装、被改变的姿势、扭曲的面部特征、过度的面部细节、过大的身体、写实的解剖结构、不必要的背景更改。
宽高比:3:4
来源与署名
原文由 @visualaiclub 发布在 X:查看原推文。 本页逐字保留原文并提供机器翻译的中文解读;版权归原作者所有。
同工具的更多提示词
- 生成前请先上传一张参考图片。 主提示词 创作一张极具戏剧性的竖版装饰艺术风格旅行海报,主题为[CITY NAME],风格借鉴1930年代豪华远洋邮轮旅行广告…@visualaiclub
- ChatGPT Image 2 vs Grok Imagine vs Gemini Nano Banana 2 💛☀️🌻 1️⃣ ChatGPT Image…@punkhuri1
- 使用上传的照片作为精确的视觉参考,创作一张高端独立艺术画册的编辑风格海报。 严格格式: - 纵向 3:4 画布 - 将构图水平划分为两个相等的 50/50 区…@AiwithLariab
- 创作一张高质量的全身 Y2K 时尚编辑风格摄影作品,主体为参考图中同一女性角色。 角色坐在洁净的白色摄影棚地面上,摆出放松、可爱的姿势,身体微微前倾,正在为一…@Wareenaa
- 创作一张高端奢华时装广告图片,主角是一位三十岁出头、气质成熟的男性模特,身着一套剪裁完美的意大利黑色西装。模特自信地站立在一个极简主义的高端时装工作室中,地面…@briepretty_728
- {"aspect_ratio": "9:16", "subject": "年轻女性,长长的波浪状棕色头发", "outfit": "鼠尾草绿色复古泡泡袖连衣裙…@hey_am_cherry