自由女神、冰激凌与美元纸币
分享链接 · https://openimages.ajiang.me/p/2059900280364556604/

camera_lens · wide-anglecomposition · centeredcomposition · full-bodycomposition · low-anglelighting · natural-daylightsubject_type · character
提示词(原文,逐字保留)
Liberty, ice cream, and dollar bills
Probably the most surreal yet oddly realistic version of the Statue of Liberty I’ve made
Created with GPT Image 2
Prompt👇
Use my uploaded portrait only as the identity reference: preserve my real facial structure, hairstyle, hairline, skin tone, and overall human likeness. Create a surreal photorealistic scene inspired by the reference image: a giant human version of me sitting like the Statue of Liberty on Liberty Island, wearing a turquoise Statue of Liberty robe and crown, holding a vanilla ice cream cone high in one hand and a realistic fish in the other hand.
Keep the composition similar to the reference: vertical 9:16 frame, low-angle tourist-photo perspective from across the water, blue sky with soft clouds, green trees and monument base below, tiny tourists at the bottom for scale, realistic daylight, natural shadows, clear environmental details. The result should look like a real high-resolution photo, not a collage. My face must be naturally integrated with the lighting and perspective, with realistic skin texture, no plastic skin, no anime style, no over-smoothing. No watermark, no logo, no text.
Negative Prompt
watermark, logo, text, Chinese characters, cat face, animal features, anime, cartoon, plastic skin, over-smoothed skin, fake face swap, pasted face, bad lighting match, distorted face, deformed hands, extra fingers, blurry, low quality, unrealistic proportions, bad perspective, messy background, artificial CGI look.
中文译文
自由女神、冰激凌与美元纸币
可能是我做过的最超现实却又异常真实的自由女神版本
使用 GPT Image 2 创作
提示词👇
仅将我上传的肖像作为身份参考:保留我真实的面部结构、发型、发际线、肤色以及整体的人类相似度。创建一个受参考图启发的超现实照片级真实场景:一个巨大的、像我本人的人类版本,以自由女神的坐姿坐在自由女神岛上,身着绿松石色的自由女神长袍并头戴王冠,一只手高举一个香草冰激凌甜筒,另一只手握着一条逼真的鱼。
保持与参考图相似的构图:竖向 9:16 画框,从水面另一侧的低角度游客照片视角拍摄,蓝色天空配以柔和的云朵,下方有绿色的树木和纪念碑基座,底部有微小的游客以体现比例感,真实的白日光照,自然的阴影,清晰的环境细节。成品应看起来像一张真实的高分辨率照片,而非拼贴画。我的面部必须与光照和视角自然融合,具有真实的皮肤质感,不要塑料感的皮肤,不要动漫风格,不要过度磨皮。无水印、无标志、无文字。
反向提示词
watermark, logo, text, Chinese characters, cat face, animal features, anime, cartoon, plastic skin, over-smoothed skin, fake face swap, pasted face, bad lighting match, distorted face, deformed hands, extra fingers, blurry, low quality, unrealistic proportions, bad perspective, messy background, artificial CGI look.
来源与署名
原文由 @Shinning1010 发布在 X:查看原推文。 本页逐字保留原文并提供机器翻译的中文解读;版权归原作者所有。
同工具的更多提示词
- // 系统:d3.js 信息图渲染引擎 const svg = d3. select("canvas").attr("width", 1920).attr("…@Gdgtify
- 关于角色设定图,我用了我在 Midjourney 中创建的一种特殊风格,稍后会与我的订阅者分享。 然后,我用 GPT Image 2 生成了故事板。 接着,我…@Artedeingenio
- 创作一幅风格化的几何立体主义体育海报,以迈克尔·乔丹(Michael Jordan)为主角,采用一个标志性的篮球姿势,灵感源自传奇的 NBA 美学与现代奢华体…@Goodmanprotocol
- 超写实生活方式时尚拼贴画,描绘一位美丽的年轻女子在黄昏金色时刻的开阔绿色公园中,柔和温暖的阳光,电影般的自然光影,韩系美学时尚氛围,穿着深酒红色超大号纽扣衬衫…@auqibhabib
- ChatGPT image 2.0 电影感竖屏首帧,场景设定在巴黎街头的黄昏时分。整体氛围为时尚浪漫喜剧风:明亮、优雅、有格调,同时扎根于真实的街头生活。 主…@Sofia_ray721
- 水印,标志,文字,动漫,卡通,娃娃脸,塑料皮肤,过度磨皮皮肤,假美颜滤镜,面部扭曲,身份不准确,复制参考脸,换脸效果,不自然的光照,糟糕的手部,多余手指,面部…@Shinning1010