OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image

真实人物与墙壁涂鸦影子交互人像

Real Person Interacting with Wall Graffiti Shadow Portrait

A polished vertical mixed-media portrait showing an identity-preserved person and a black hand-drawn doodle-shadow on the wall beside them, sharing a newly invented playful pose.

真实人物与墙壁涂鸦影子交互人像 — AI image creation
真实人物与墙壁涂鸦影子交互人像

提示词 / Prompt

中文提示词

使用已提供的上传角色参考图作为严格的身份与装束参考，保持人物的面部身份、面部比例、眼睛形状、鼻子、嘴唇、肤色、发型、发色、可见发饰、整体可识别气质，以及参考图中的服装和造型；不要硬编码参考图中不存在的蓝发、发夹、校服或其他特征，所有细节必须直接依据参考图。创作一张高质量纵向混合媒介人像，展示上传角色的写实全身或近全身版本，以及旁边干净墙面上同一角色的黑色手绘涂鸦影子。涂鸦影子必须是直接画在墙上的黑色线稿，不是写实第二个人、正常物理影子、全彩动漫角色或怪物。它通过发型轮廓、配饰、服装轮廓和姿势结构与真人相似，但比真人更夸张、更滑稽、更混乱，可配少量漫画运动线、爱心、星星、闪光和标记。每次生成都发明一个新的顽皮姿势，不能固定或重复指向、手指枪、镜像站立姿势；真人做自然可爱的现实版本，表情略困惑、轻微尴尬、像在不情愿地配合，涂鸦影子则像玩得过于开心。姿势可以是偶像、舞蹈、侧倾、大爱心、眨眼、双手贴脸、哒哒、惊讶、原地跑、踮脚、躲猫猫、平衡或搞怪胜利等，但不固定为菜单。使用 4:5 或 9:16 纵向构图，真人置于一侧，涂鸦影子在旁边墙上，保留足够空间展示完整动作与两者联系。背景为白色、米色或浅灰色的干净室内墙和地板，可有柔和自然光斑；光线明亮、自然、精致，人物面部清晰。避免凌乱背景、姿势不匹配、扭曲身体、多余手指、文字、标志、水印或明显 AI 瑕疵。

English prompt

Use the supplied uploaded character reference as the strict identity and outfit reference. Preserve the subject's facial identity, facial proportions, eye shape, nose, lips, skin tone, hairstyle, hair color, visible hair accessories, recognizable overall vibe, and the clothing and styling shown in the reference. Do not hardcode blue hair, a specific hair clip, a school uniform, or any other feature absent from the reference; infer all such details directly from the image. Create a high-quality vertical mixed-media portrait showing a realistic full-body or nearly full-body version of the uploaded character and a black hand-drawn doodle-shadow of the same character on the clean wall beside them. The doodle-shadow must be black line art drawn directly on the wall, not a realistic second person, normal physical shadow, full-color anime character, or monster. It should resemble the real person through hairstyle silhouette, accessories, outfit silhouette, and pose structure, while being more exaggerated, silly, energetic, and chaotic; add a few manga-like motion lines, hearts, stars, sparkles, and comic marks if useful. Invent a fresh playful mischievous pose for every generation; never fix or repeat pointing, finger-gun, or mirrored standing poses. The real person performs a cute natural realistic version with a slightly confused, mildly embarrassed expression, as if reluctantly playing along; the doodle-shadow looks as if it is having far too much fun. Loose inspiration may include idol, dance, sideways lean, big heart, wink, hands-near-cheeks, ta-da, mock surprise, running-in-place, tiptoe, peekaboo, balance, or goofy victory poses, but do not use a fixed menu. Use a 4:5 or 9:16 vertical composition, placing the real person on one side and the doodle-shadow on the adjacent wall with enough space for the full pose and clear connection between them. Use a clean white, cream, or pale-gray indoor wall and floor, with optional soft sunlight patches; keep the lighting bright, natural, polished, and the face clearly visible. Avoid cluttered backgrounds, mismatched poses, distorted bodies, extra fingers, text, logo, watermark, or obvious AI artifacts.

查看 API 模型

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