基于照片参考生成动漫肖像
将一张简单的坐姿参考照片转换为动漫风格的特写肖像，并更换服装、发色及绿植背景。
gpt-image-2 · 个人资料 / 头像 · 动漫 / 漫画 · 社交媒体帖子 · 角色
中文提示词
以 REFERENCE_0 作为姿势和构图基础，将照片中的坐姿女性转换为精致的动漫风格角色肖像。保持面部匿名化处理（使用大的柔和矩形模糊块），并将构图从侧视的咖啡馆/餐桌场景改为正面的半身特写。将原有的写实室内环境、桌子、马克杯、书本和椅子替换为明亮的绿植室内花园/窗户背景。为角色重新设计发型，使用 {argument name="hair color" default="淡冰蓝色"} 的短波波头，并让她穿上华丽的 {argument name="outfit style" default="黑白银配色哥特偶像风无袖上衣"}，搭配蕾丝肩部细节、丝带系带、装饰性纽扣和结构化拼接设计。采用清晰、高质量的日本动漫插画风格，具备干净的线条、柔和的灯光、生动的绿植、浅景深效果以及竖屏手机构图。在保留参考图中平静坐姿氛围的同时，使最终图像呈现出完整的角色插画质感，而非照片效果。
原始提示词
Using REFERENCE_0 as the pose and composition base, transform the seated woman into a polished anime-style character portrait. Keep the face anonymized with a large soft rectangular blur block, but change the framing from a side-view café/table scene into a close-up front-facing bust portrait. Replace the original realistic indoor room, table, mug, book, and chair with a bright leafy indoor garden/window background. Restyle the character with {argument name="hair color" default="pale icy blue"} short bob hair, and dress her in an ornate {argument name="outfit style" default="black, white, and silver gothic idol-style sleeveless blouse"} with lace shoulder details, ribbon tie, decorative buttons, and structured paneling. Use crisp high-quality Japanese anime illustration style with clean line art, soft lighting, vivid greenery, shallow depth of field, and a vertical smartphone composition. Preserve the calm seated mood from the reference while making the final image look like a finished character illustration rather than a photo.