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

米色沙发上的暖光自拍

Warm-Lit Selfie on a Beige Sofa

采用清晰文字姿势的居家沙发自拍；成人身份参考与铂金金色发型适配均已注明。

米色沙发上的暖光自拍 — AI image creation
米色沙发上的暖光自拍

提示词 / Prompt

中文提示词

创作一张 3:4 写实摄影肖像，主体是一位明确成年的女性，使用附件身份参考呈现其可辨认的面部身份。她将头发随意挽成铂金金色的松散发髻，几缕发丝垂在脸侧；这是对参考中可见深色头发所作的明确发型适配。她穿浅蓝色圆领背心和米色抽绳睡裤。她放松地斜倚坐在一张大型、有纹理的米色组合沙发上：躯干略向右转，头部朝向镜头，右臂向镜头伸出形成手臂长度的自拍，左肩靠在沙发靠垫上。采用略高的手机自拍视角向下拍摄，保留自然近距离透视；面部和上半身清晰，背景清晰度柔和递减。她身后有炭灰色天鹅绒靠垫和有纹理的米色靠垫，墙上有一幅部分入镜、类似老式汽车车轮的黑白相框照片。画面右侧放置一盏黑色灯罩的暖光台灯。柔和温暖的主光从左前方照亮脸部和上半身，背景保持昏暗舒适。妆容精致，睫毛分明，修容自然，唇部有光泽；她露出灿烂愉快的笑容并直视镜头。采用高保真社交媒体人像摄影质感，呈现真实皮肤纹理、自然面料褶皱和清晰织物细节。保持来源指定的 3:4 构图。来源原文在写出 DWPose 后截断于“pu”；以上姿势已由完整文字明确描述，不声称提供或运行 DWPose/ControlNet 输入。来源中的 8K 等质量表述代表效果意图，不保证实际输出分辨率。

English prompt

Create a photorealistic 3:4 portrait of a clearly adult woman, using the attached identity reference for her recognizable facial identity. Her hair is styled in a casual platinum-blonde loose updo with a few face-framing strands; this is a disclosed hair adaptation from the reference's visible dark hair. She wears a light-blue scoop-neck tank top and cream drawstring pajama pants. Show a relaxed seated recline on a large textured cream sectional sofa: torso angled slightly right, head turned toward the camera, right arm extended toward the lens for an arm's-length selfie, and left shoulder resting against a sofa cushion. Use a slightly elevated phone-camera angle looking down, with natural close-range perspective and sharp focus on face and upper body, falling off gently toward the background. Behind her are charcoal velvet and textured beige throw pillows and a partly visible framed black-and-white photograph resembling a vintage car wheel. Place a warm glowing lamp with a black shade on the right. Soft flattering warm indoor light falls from the front-left across her face and upper body; keep the background dim and cozy. Give her polished makeup, defined lashes, subtle contour, glossy lips, a broad happy smile and direct eye contact. Use high-fidelity social-media portrait photography, realistic skin texture, natural fabric folds and clearly rendered textile surfaces. Keep the source's stated 3:4 composition. The source text ends mid-field after naming DWPose and the fragment 'pu'; the pose above is fully specified in text, and no DWPose/ControlNet input or execution is claimed. Source quality labels such as 8K describe intent, not guaranteed output resolution.

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