Use the uploaded character reference image as the strict identity and outfit reference. Preserve the reference character’s: face identity facial proportions eyes nose lips skin tone hairstyle hair color visible hair accessories overall recognizable vibe outfit and styling Infer all identity, hairstyle, accessories, and clothing only from the uploaded reference. Do not add traits not present. Create a high-quality mixed-style vertical portrait showing: 1. a realistic full-body version of the uploaded character/person 2. a black hand-drawn doodle-shadow of the same character on the wall beside them Core concept: The real person and their doodle-shadow share a cute, mischievous pose. The real person performs a realistic version, while the doodle-shadow exaggerates it with chaotic cartoon energy. The mood is playful, stylish, funny, and social-media-friendly. Real person: realistic and photogenic keep the same outfit and styling from the reference cute, slightly embarrassed, playful expression natural body language actively joins the pose Doodle-shadow: black sketch line-art drawn on the wall not a real person or normal shadow resemble the person through hairstyle, accessories, outfit silhouette, and pose more energetic and exaggerated optional manga motion lines, hearts, stars, sparkles, and comic effects Random mischievous pose rule: Generate a different playful pose every time. The real person and doodle share the same pose idea but not identical execution. Avoid repeated pointing, finger-guns, stiff standing, mirrored poses, or fixed compositions. Examples include: idol pose silly dance ta-da mock surprise playful tiptoe sneaking away peekaboo dramatic overreaction balance pose victory pose shy troublemaker overconfident pose Composition: vertical 4:5 or 9:16 full or nearly full body real person on one side doodle-shadow on a clean wall beside them similar height clear interaction Background: simple indoor studio or minimal room white, cream, or pale gray wall clean floor uncluttered soft sunlight or gentle wall shadow allowed Lighting: soft natural studio lighting bright, clean, playful face clearly visible Style quality: realistic photography black doodle-shadow faithful identity and outfit strong resemblance clean composition polished mixed-media portrait no obvious AI artifacts Negative prompt: Hardcoded hair/accessories/outfits not in reference, unrelated outfit changes, realistic second person, normal reflection/shadow, horror or monster shadow, full-color cartoon/anime human, weak resemblance, unrelated sketch, cluttered background, repeated or fixed poses, stiff body, bad hands, extra limbs/fingers, distortion, text, watermark, logo. 使用上传的角色参考图作为严格的身份和服装参考。 保留参考角色的: 面部身份 面部比例 眼睛 鼻子 嘴唇 肤色 发型 发色 可见的发饰 整体可辨识的氛围 服装和造型 仅从上传的参考图中推断所有的身份、发型、配饰和服装。不要添加不存在的特征。 创建一个高质量的混合风格纵向肖像,展示: 1. 上传的角色/人物的写实全身版本 2. 墙上旁边同一个角色的黑色手绘涂鸦影子 核心概念:真人及其涂鸦影子共享一个可爱、调皮的姿势。真人表演写实版本,而涂鸦影子以混乱的卡通能量夸大它。色调是俏皮、时尚、有趣且适合社交媒体的。 真人: 写实且上镜 保持参考图中的相同服装和造型 可爱、略显尴尬、俏皮的表情 自然的肢体语言 积极加入姿势 涂鸦影子: 画在墙上的黑色素描线稿 不是真人或正常的影子 通过发型、配饰、服装轮廓和姿势与本人相似 更具活力且夸张 可选的漫画动作线、爱心、星星、闪烁和漫画效果 随机调皮姿势规则:每次生成不同的俏皮姿势。真人和涂鸦共享相同的姿势创意,但执行方式不同。避免重复的指点、手指枪、僵硬的站立、镜像姿势或固定的构图。 示例包括: 偶像姿势 傻傻的舞蹈 哒哒(亮相) 假装惊讶 俏皮的踮脚尖 偷偷溜走 捉迷藏 戏剧性的过度反应 平衡姿势 胜利姿势 害羞的麻烦制造者 过度自信的姿势 构图: 纵向 4:5 或 9:16 全身或近乎全身 真人在一侧 侧面干净墙上的涂鸦影子 高度相似 清晰的互动 背景: 简单的室内工作室或极简房间 白色、奶油色或浅灰色墙壁 干净的地板 整洁 允许柔和的阳光或柔和的墙壁阴影 光照: 柔和的自然工作室光照 明亮、干净、俏皮 面部清晰可见 风格质量: 写实摄影 黑色涂鸦影子 忠实的身份和服装 强烈的相似度 干净的构图 精致的混合媒体肖像 没有明显的 AI 瑕疵 负面提示词:硬编码参考图中没有的发型/配饰/服装,无关的服装变化,写实的第二个人,正常的反射/阴影,恐怖或怪物影子,全彩卡通/动漫人类,相似度弱,无关的素描,杂乱的背景,重复或固定的姿势,僵硬的身体,坏手,多余的肢体/手指,扭曲,文字,水印,标志。
{ "direction": "Not an advertising photo, but the kind of ultra-close-up vertical selfie taken by a friend holding a phone to their face in a dim cafe. An adult Japanese woman in her 20s. The character occupies about 85-90% of the screen,