OpenZhenZhen AI 提示词画廊 · GPT Image 2.6 · image

保留人物原貌的青紫霓虹边缘光照片编辑

Portrait photo edit with violet neon rim lighting, original features preserved

以同一完整 ref-10a-v1 横幅照片为底图，仅改变照明；人物及所有其他可见内容与原画幅均保持不变。

保留人物原貌的青紫霓虹边缘光照片编辑 — AI image creation
保留人物原貌的青紫霓虹边缘光照片编辑

提示词 / Prompt

中文提示词

将完整的 identity_reference ref-10a-v1 原图本身编辑为一张静态照片；同一整幅图既是编辑底图，也是面部身份参考。保留原有完整横幅画面和准确构图、同一成年人物、面部及面部结构、身体与比例、姿势、表情、发型、服装、皮肤纹理、背景、物件、场景几何结构、透视和现有细节。唯一允许改变的是照明。维持深色低调的基础观感并保留自然可辨细节，沿原有面部和发丝轮廓加入鲜明但受控的青色与洋红色霓虹边缘高光，在较深阴影中形成彩色点缀。来源未指定霓虹色，因此青色与洋红色是本配方明确选定的默认方案。避免让彩光变成对皮肤、服装或背景表面的重新着色；不要标准化预期的暗部曝光或色彩，不做自动曝光校正，也不添加美颜光。保留原始取景和画幅，不裁切、不扩图、不重新构图。不美化，不重塑面部或身体，不改变姿势、表情、发型或服装，不加物件、规定照明以外的光晕效果、文字、标志、水印或动态镜头。

English prompt

Edit the complete attached identity_reference ref-10a-v1 itself as one still photograph; this same full image is both the edit base and face-identity reference. Preserve the original full horizontal frame and exact composition, same adult person, face and facial structure, body and proportions, pose, expression, hair, clothing, skin texture, background, objects, scene geometry, perspective and existing detail. Lighting is the only permitted change. Keep a dark, low-key base appearance with readable natural detail, and add vivid but controlled cyan and magenta neon rim highlights along the existing face and hair edges, creating colored accents against deeper shadows. The cyan-and-magenta pairing is a selected default because the source leaves accent colors open. Keep colored light from becoming a recoloring of skin, clothing or background surfaces; do not normalize the intentional dark exposure or color, apply automatic exposure correction, or add beauty lighting. Preserve the original framing and aspect without crop, expansion or reframing. No beautification, face or body reshaping, pose or expression changes, hair or clothing edits, extra objects, glow effect beyond the specified illumination, text, logo, watermark or motion.

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