OpenZhenZhen AI Prompt Gallery · gpt-image-2.5 · image

Porcelain Skin Watercolor Illustration

Porcelain Skin Watercolor Illustration — explore this gpt-image-2.5 creation on OpenZhenZhen, with its original prompt and a reusable generation example.

Porcelain Skin Watercolor Illustration — AI image creation
Porcelain Skin Watercolor Illustration

Prompt

Chinese prompt

将配置人物参考ref-10a-v1中的成年主体转化为特写时尚编辑插画，同时保留其可识别身份、脸型、眼睛、鼻子、嘴部、自然比例与发型。不要承诺像素级完全复刻；应在艺术化处理后仍清楚保持人物相貌特征。融合高端插画写实、精致墨线与柔和水彩笔触。细致描绘面部，以瓷质肌肤纹理、轻柔腮红和富有表现力的闪亮眼睛呈现。采用动态近景肖像构图，背景保留清爽纯白留白。叠加克制的建筑轮廓、几何结构网格与技术制图线条，使其轻盈并从属于面部。让飘逸的发丝与织物边缘渐变消融为透明粉彩碎片和颜料飞溅。使用柔和蓝色、暖黄色、淡粉色和浅灰色，并控制饱和度。以精致准确的墨线轮廓和疏落的建筑及几何制图线条，平衡精准素描线条与松散氛围感颜料效果。整体呈现优雅的电影感编辑插画与现代时尚艺术气息，梦幻而锐利，极简却细节丰富。不增加其他人物、文字、标志或水印。

English prompt

Transform the adult subject in configured reference ref-10a-v1 into a close-up fashion-editorial illustration, preserving their recognizable facial identity, face shape, eyes, nose, mouth, natural proportions and hairstyle. Do not promise a pixel-identical face; keep the likeness clear through the artistic treatment. Blend high-end illustrative realism with delicate ink lines and soft watercolor strokes. Render detailed facial features with porcelain-like skin texture, a gentle blush and expressive sparkling eyes. Use a dynamic close-up portrait composition on a crisp white negative-space background. Overlay restrained architectural outlines, geometric construction grids and technical drawing marks, keeping them light and secondary to the face. Let flowing hair and fabric edges dissolve into transparent pastel shards and paint splatters. Use soft blue, warm yellow, pale pink and light gray with controlled saturation. Balance delicate precise ink contours with sparse architectural and geometric drafting lines against loose atmospheric paint effects. Create an elegant cinematic editorial illustration with a modern fashion-art feel: dreamy yet sharp, minimalist yet rich in detail. No extra people, text, logos or watermark.

View API model

Explore more AI image and video prompts · Try image generation · Try video generation · Create on the AI canvas