OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image
原图编辑增加真实同卵双胞胎姐妹
Original Edit: Adding Real Identical Twin Sisters
来自 OpenNana 的图像提示词，已按本地图像分类和站点 gpt-image-2.5 归类整理。 使用时需上传人物参考图；本流程使用统一参考人物生成示例。
提示词 / Prompt
中文提示词
以用户上传的参考图作为主要可见人物的面部身份依据。在不与身份参考冲突的情况下，保留来源的场景、造型、服装、姿势、构图、光线和视觉风格。 以我上传的原图作为唯一编辑基础进行局部修改，不要重新生成或重新设计整个画面。最大程度保留原人物、背景、构图、镜头距离、透视、机位、光线、阴影、色调、空间结构和原有摄影风格。 保留原图人物本人以及她原来的脸、年龄感、发型、发色、肤色、身材比例、体型、服装、位置和主要姿势，不要换脸、不要重新美化、不要改变原人物身份。 在原人物旁边自然增加另一位女性，设定为她的同卵双胞胎姐妹。新增人物必须基于原人物本人，保持相同的脸型、五官结构、年龄感、肤色、发型、发色、体型、身体比例和整体气质，两人一眼看上去必须是真实双胞胎，而不是两个相似模特。 不要简单复制粘贴。新增人物可以在头部角度、眼神、手部动作和身体朝向上有非常轻微的变化，使双胞胎关系更加真实自然。 两人距离较近，表现轻微依偎或轻微拥抱。可以让新增人物轻轻靠向原人物，肩膀贴近，手臂轻轻搭在肩部、上臂或腰侧，动作柔和、克制、自然，不要过度拥抱，不要夸张肢体动作。 新增人物的服装必须保持与原人物相同的设计、剪裁、材质、结构和风格，只改变颜色。根据原服装颜色自动选择柔和的视觉补色，例如蓝色对应橙色或琥珀色、粉色对应薄荷绿、紫色对应黄绿色。补色要低饱和、协调、时尚，不要荧光或过度抢眼。 新增人物必须完全匹配原图光线，包括主光方向、阴影、皮肤高光、头发高光、衣服反光、环境反射和接地阴影。两个人的手臂、肩膀、头发、衣服和身体接触必须符合真实物理关系，不允许穿模或错误重叠。 最终效果应像这对双胞胎从拍摄开始就同时存在于原场景中。保持真实摄影感、自然皮肤纹理、真实发丝、正确人体结构和真实透视。避免动漫脸、CGI、塑料皮肤、复制感、镜像感、错误手指、错误四肢、人体融合、错误透视、背景大改、服装款式大改和明显 AI 合成痕迹。
English prompt
Use the uploaded reference image as the facial identity reference for the main visible person. Keep the source scene, styling, clothing, pose, composition, lighting and visual style unless they conflict with the identity reference. Use my uploaded original image as the sole editing basis for local modification, do not regenerate or redesign the entire screen. Maximize retention of the original person, background, composition, lens distance, perspective, camera angle, light, shadow, tone, spatial structure, and original photography style. Retain the original person herself and her original face, sense of age, hairstyle, hair color, skin tone, body proportions, body type, clothing, position, and main posture; do not swap faces, do not re-beautify, do not change the identity of the original person. Naturally add another woman next to the original person, set as her identical twin sister. The newly added person must be based on the original person herself, maintaining the same face shape, facial structure, sense of age, skin tone, hairstyle, hair color, body type, body proportions, and overall temperament; the two must look like real twins at first glance, not two similar models. Do not simply copy and paste. The newly added person can have very slight changes in head angle, gaze, hand movements, and body orientation to make the twin relationship more real and natural. The two are close together, showing a slight snuggle or a slight hug. Let the newly added person lean gently toward the original person, shoulders close, arms gently resting on the shoulder, upper arm, or side of the waist; movements should be soft, restrained, and natural, no excessive hugging, no exaggerated body movements. The clothing of the newly added person must maintain the same design, cut, material, structure, and style as the original person, only changing the color. Automatically select soft visual complementary colors based on the original clothing color, such as blue corresponding to orange or amber, pink corresponding to mint green, purple corresponding to yellow-green. Complementary colors should be low-saturation, coordinated, and fashionable, not fluorescent or overly eye-catching. The newly added person must completely match the original image's lighting, including main light direction, shadows, skin highlights, hair highlights, clothing reflections, environmental reflections, and grounding shadows. The contact between the two people's arms, shoulders, hair, clothes, and bodies must comply with real physical relationships; no clipping or incorrect overlapping allowed. The final effect should look as if these twins existed in the original scene from the start of the shoot. Maintain a realistic photography feel, natural skin texture, real hair strands, correct human body structure, and real perspective. Avoid anime faces, CGI, plastic skin, copy-paste feel, mirror-image feel, wrong fingers, wrong limbs, body fusion, wrong perspective, major background changes, major clothing style changes, and obvious AI synthesis traces.
探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作
相关作品
- 万圣节甜酷暗黑洛丽塔高级定制服装
H-LOOK · Halloween Collection 的橙黑与奶油白万圣节造型：束身胸衣、蓬袖、不对称多层蛋糕短裙、荷叶边、缎面蝴蝶结、南瓜吊饰及厚底高跟鞋，以无脸服装展示形式呈现。 | An orange, black, and cream H-LOOK · Halloween Collection outfit featuring a corseted bodice, puff sleeves, an asymmetrical layered cupcake miniskirt, ruffles, satin bows, pumpkin charms, and platform he
- 海边观景点白帽与白外套手机抓拍
一位二十岁出头的成年女性在海边墙旁拍摄竖幅旅行快照，戴白色棒球帽，穿罗纹背心和半滑落的轻便白外套；左上方强烈日光与海面冷调反光形成对比。 | An adult woman in her early 20s is captured in a vertical travel snapshot by a wall at a seaside viewpoint, wearing a white baseball cap, ribbed tank and lightweight jacket slipping off her shoulders, with strong upper-left sun a
- 镜前红黑格裙与大型镰刀道具角色扮演
一位二十多岁的成年女性穿黑色外套与红黑格裙角色扮演服，站在全身镜前，一手持大型镰刀道具、一手拿手机；房间柔光与红色装饰形成对比。 | An adult woman in her 20s wears a black coat and red-and-black plaid cosplay outfit before a full-length mirror, holding a large scythe prop in one hand and a phone in the other, with soft room lighting and red accents.
- 麦当劳窗边双人早餐生活抓拍
两位成年年轻东亚女性在麦当劳落地窗边的高桌旁托腮交谈，桌上摆着汉堡、薯条和饮料；透过玻璃可见城市街道，室内日光交融，呈现真实随性的生活瞬间。固定面部参考仅用于左侧女性。 | Two adult young East Asian women chat with their chins resting on their hands at a high table by a McDonald's floor-to-ceiling window, with burgers, fries and drinks on the table and the city visible outside in mi
- 冷白桃花粉双马尾梦幻面部特写
采用 1:1 手机室内近摄，以冷白、乌黑和桃花粉的低饱和柔雾色调刻画成年女性的双马尾刘海、垂眸眼妆与银色耳饰，构成安静梦核感的局部面部特写。固定面部参考仅提供身份线索。 | A 1:1 indoor phone close-up uses a low-saturation, mist-soft palette of cool white, jet black and peach blossom pink to depict an adult woman's twin tails, downcast eye makeup and silver earrings in a quiet dreamc
- 东京晨蓝便利店街头早餐纪实写真
上午 9:03 的日本街角便利店外，一位 25 岁成年东亚女性倚着玻璃墙，手拿冰美式与饭团，像刚好被捕捉到的日常瞬间；画面以冷白晨光和低饱和灰蓝胶片色调呈现。固定面部参考仅提供身份线索。 | At 9:03 AM outside a Japanese corner convenience store, a 25-year-old adult East Asian woman leans against the glass with an iced Americano and rice ball, captured in a candid everyday moment with cool mo