OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image
古装女性九宫格连续穿衣过程
Nine-Grid Sequence of a Woman in Traditional Dress Getting Dressed
用途：古装穿衣过程九宫格。保留同一成年东方女性、橙红与米棕唐风服饰从内衣到完整造型的九步连续变化、梳妆台与铜镜空间及DeepBlue签名。
提示词 / Prompt
中文提示词
以 ref-10a-v1 作为成年女性的面部身份参考，保留来源造型与场景。 以提供的 ref-10a-v1 参考图仅锁定同一成年东方女性的面部身份、发型、身体比例和完整服装设计，不复制参考图白色背景或角色设定板排版。制作9:16竖版海报，严格3×3等大网格和窄间距。同一成年女性在古代闺房连续完成九个穿衣步骤：(1) 传统素色内衣站在木衣架旁，头发披散；(2) 穿上橙红贴身内层；(3) 穿橙红长裙；(4) 套入浅米棕半透明宽袖外袍；(5) 整理外袍、衣襟与层叠裙摆；(6) 在梳妆镜前系上参考图腰带和金属装饰；(7) 调整袖口与裙摆；(8) 坐在闺房软榻边穿参考图古鞋；(9) 完成双髻和全套橙红/米棕唐风服装，在古代铜镜前自然站立。每格保留前一步已完成衣层，只改变当前步骤；脸、发色、比例、色彩、半透明材质、袖型、裙型、纹样和配饰始终一致。闺房含木格窗、红纱、衣架、梳妆台、铜镜、木家具、帘幕、花枝及暖烛火。真实编辑类古装摄影，自然皮肤、发丝、织物和半透明层，九格光影透视协调；不要现代物件、现代服饰、重复人物、缺失衣层、额外文字或水印，右下角仅保留小型深蓝圆点与白色手写 DeepBlue 签名。
English prompt
Use ref-10a-v1 as the facial identity reference for the adult woman. Preserve source styling and scene. Use the supplied ref-10a-v1 reference image only for the same adult Oriental woman's facial identity, hair, body proportions and complete costume design; do not copy its white background or character-sheet layout. Create a 9:16 vertical poster with a strict 3×3 grid, equal cells and narrow spacing. The same adult woman completes nine sequential dressing stages in an ancient woman's boudoir: (1) plain traditional inner garments beside wooden rack, hair loose; (2) red-orange fitted inner top; (3) red-orange long skirt; (4) step into translucent light beige-brown wide-sleeved outer coat; (5) finish robe and adjust layered hem; (6) tie the reference waist belt and metal decorations at vanity mirror; (7) adjust sleeves and hem; (8) sit on couch putting on reference footwear; (9) complete double-bun hair, full red-orange/beige-brown Tang-style outfit, and stand naturally before an ancient bronze mirror. Every cell retains the prior completed layers and changes only the stated step; face, hair color, proportions, palette, fabric transparency, sleeves, skirt, patterns and accessories stay consistent. Ancient boudoir: wooden lattice window, red gauze, rack, vanity, bronze mirror, wooden furniture, curtains, flowers and warm candle/oil light. Photorealistic editorial costume photography with natural skin, hair, textiles and translucent layers; coherent shadows and perspective across cells, no modern objects, modern clothing, duplicated people, missing layers, text or watermark except a small DeepBlue blue circular mark with white handwriting in a consistent lower corner.
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
相关作品
- 十二宫格人物发型设计大赏
以已配置的人像作为角色身份参考，设计十二种不同发型，并分别搭配服装与姿势，以十二格展示；为每格提供造型适配度评分。 Use the configured portrait as the character identity reference to create twelve hairstyle variations with coordinated clothing and poses in a twelve-cell grid, each with an individual styling-suitability score.
- 夏日八套穿搭自由拼贴画
以固定女性肖像作为脸部身份参考，创作八套全身夏季穿搭的自由式双排拼贴；保持人物脸部、比例、尺度和机位一致，不在画面文字中写身高。 Create a freeform two-row editorial collage of eight full-body summer looks, using the configured portrait as the preview face identity and keeping the face, proportions, scale and camera distance consistent without printing a height.
- 超写实时尚杂志封面大片
以已配置的固定女性肖像为身份参考，创作坐在雕塑沙发上的超写实 VOGUE 高级时尚封面，保留指定服装、妆容、影棚光线与双语封面文案。 Create a hyper-realistic high-fashion VOGUE cover using the configured fixed woman portrait as identity reference, preserving the specified editorial styling, lighting and exact bilingual cover copy.
- 极简米色棚拍人物特写
在柔和米色影棚中制作九格杂志风肖像，九个镜头保持同一参考人物身份，同时呈现指定服装、不同取景与内省氛围。 Create a nine-frame editorial portrait grid in a soft beige studio, preserving the same reference identity across distinct specified shots, wardrobe and introspective mood.
- 质感拉满的高端商业美女广告
以已核验的当前固定女性肖像保持面部身份，制作可替换产品与品牌氛围的高端商业广告；默认产品为咖啡、氛围为奢华，不依赖特定品牌或产品图。 Create a premium commercial ad using the verified configured woman’s portrait for face identity, with replaceable product and brand mood; default to coffee and luxury without requiring a specific brand or product image.
- 高级极简人像海报设计
以当前固定人像作为本次原始人物照片，制作3:4竖版高级极简肖像海报；保留人物可见身份、发型、发色、服装和姿势，并用可替换的三行英文默认文案。 Use the configured portrait as this preview’s original photo for a 3:4 minimalist poster, preserving visible identity and appearance with three replaceable English copy lines.