OpenZhenZhen AI 프롬프트 갤러리 · gpt-image-2.5 · image
9:16 인물 포즈 참고와 숨겨진 옆얼굴 그라데이션 배경
9:16 Character Pose Reference with Hidden Profile Gradient Background
‘9:16 인물 포즈 참고와 숨겨진 옆얼굴 그라데이션 배경’은(는) gpt-image-2.5로 만든 작품입니다. OpenZhenZhen에서 원본 프롬프트와 생성 예시를 살펴보고 새로운 창작에 활용해 보세요.
프롬프트
중국어 프롬프트
只生成1张完整的9:16竖版4×4姿势图，共16个独立编号画面。所有画面使用同一位一位成年东亚女性（英文对照：one adult East Asian woman），明确为成年人，身份固定为ref-10a-v1；默认主体是一位成年东亚女性。16格中的身份、面部、外貌、体型与发型保持一致。她穿不透明黑色高领合身运动上衣与不透明全长紧身裤（英文对照：an opaque black high-neck fitted athletic top and opaque full-length leggings），采用不透明、完整遮盖的编辑人像训练服；各格服装颜色与设计一致。头发、衣摆和褶皱可随动作自然变化。主体始终是第一视觉重点。 最高优先级：16格必须由身体姿势结构本身区分，不能仅靠镜头角度、人物朝向、表情或裁切制造差异。严格按编号顺序执行以下姿势，不交换、不重复、不遗漏：01后方跪坐，双膝着地、臀部靠近脚跟、上身直立并背向镜头；02双膝跪地，一腿向侧方伸展、另一腿弯曲，上身直立；03双手双膝着地形成斜向四点支撑，一腿向侧前方伸展；04完整俯卧，前臂支撑上身，一腿向后上方弯曲抬起，另一腿伸展；05完整侧卧，一只手臂支撑上身，双腿伸展；06侧坐，一腿向前、一腿收拢，身体后倾并用一只手撑地；07正面跪坐，双膝着地、双腿向两侧大幅展开，双手自然放置；08坐姿，一腿收拢、一腿伸展，一臂上举、另一手维持平衡；09俯卧，双腿向后伸、双臂向前伸直；10后方或斜后方跪姿，一腿向侧后方伸展，上身直立；11仰卧，背部贴地、双膝向上弯曲、双脚靠近地面；12极低位后方四点支撑，双手双脚着地、臀部抬高形成紧凑倒V形；13坐姿，双腿在身前收拢交叠、双臂环抱；14侧卧，一手托头、双腿同时弯曲收拢；15仰卧，双腿抬离地面并向上弯曲；16侧坐，双腿完全折叠至身体一侧，身体后倾并用一只手撑地。每种姿势的身体结构都必须清楚且彼此不同。 严格按以下宫格顺序排列：01 02 03 04 / 05 06 07 08 / 09 10 11 12 / 13 14 15 16。每格只出现同一主体在对应姿势中的一个完整画面。手、脚、膝、肘及支撑接触关系完整入镜且清晰。格间边界明确，不重叠、不串格、不交换、不遗漏、不重复。统一Editorial人像摄影语言、人物比例、服装与背景逻辑。 根据主体发色和服装，从不超过两种核心色相中提取低饱和、低对比度的极简渐变：低饱和炭灰与暖石色（英文对照：muted charcoal and warm stone）。只用大面积纯净色彩和缓慢渐变，不添加环境、道具、纹理、复杂光影或装饰元素。16格共享同一连续渐变平面并保留大量负空间，背景不得形成具体场景。 隐藏视觉：在16格背景中连续形成同一主体的45度侧脸轮廓，依据其固定身份与发型特征生成，低对比地拆分为16个自然衔接的片段。整张图合看时可辨认完整轮廓，单格只出现局部抽象明暗形态。仅用极弱明暗变化，不画眼、鼻、嘴或具体五官。这只是同一人物的无五官背景图形，不是第二个人物。不得覆盖姿势主体、削弱人物清晰度、形成水印或与宫格争夺注意力。 视觉层级：第一层为16个清晰姿势；第二层为安静极简渐变；第三层为几乎不可见的连续侧脸轮廓。最终只输出一张完整9:16竖版4×4图片。不要添加标题、说明文字、logo、水印、杂志排版文字或其他字样。所有姿势保持完整着装、非露骨，适合作为编辑人像姿势参考。
영어 프롬프트
Create exactly one complete 9:16 vertical 4×4 pose-map image with sixteen independent numbered frames. Use one adult East Asian woman as the same clearly adult person in every frame (Chinese description: 一位成年东亚女性), with identity fixed to ref-10a-v1. The default subject is one adult East Asian woman. Keep her identity, face, appearance, body type and hair consistent in all cells. Dress her in an opaque black high-neck fitted athletic top and opaque full-length leggings (Chinese counterpart: 不透明黑色高领合身运动上衣与不透明全长紧身裤), an opaque, fully covering editorial training outfit; keep its colors and design consistent. Hair, hems and wrinkles may move naturally with each pose. The subject remains the primary visual focus. Highest priority: the physical pose structure—not camera angle, orientation, expression or crop alone—must distinguish all sixteen frames. Execute these numbered poses exactly in order, without swapping, repeating or omitting: 01 rear kneeling, knees down and hips near heels, torso upright facing away; 02 kneeling with one leg extended sideways and the other bent, torso upright; 03 diagonal four-point kneel on both hands and knees with one leg extended toward the side-front; 04 fully prone, forearms supporting torso, one leg bent and lifted back/up, other extended; 05 fully side-lying with one arm supporting the torso and legs extended; 06 side-sitting, one leg forward and one tucked, leaning back on one supporting hand; 07 front-facing kneel with knees down and legs opened wide to either side, hands natural; 08 seated with one leg tucked, one extended, one arm raised and the other balancing; 09 prone with legs back and both arms straight forward; 10 rear/diagonal-rear kneel with one leg extended sideways/back, torso upright; 11 supine, back flat, knees bent upward and feet near the floor; 12 ultra-low rear four-point stance on hands and feet, hips raised in a compact inverted V; 13 seated with legs crossed/tucked in front and arms hugging them; 14 side-lying with one hand supporting the head and both legs bent/tucked; 15 supine with both legs lifted and bent upward; 16 side-sitting with both legs folded to one side, leaning back on one hand. Keep each pose anatomically clear and distinct. Use this strict grid order: 01 02 03 04 / 05 06 07 08 / 09 10 11 12 / 13 14 15 16. Each cell contains only one complete depiction of the same subject in its corresponding pose. Keep hands, feet, knees, elbows and support contacts in-frame and readable. Separate all cell boundaries cleanly; no overlap, bleed, swaps, omissions or repetitions. Unify editorial portrait-photography language, scale, outfit and background logic. Create a minimalist, low-contrast, low-saturation gradient from no more than two core hues derived from the subject's hair and outfit: muted charcoal and warm stone (Chinese counterpart: 低饱和炭灰与暖石色). Use broad clean color areas and slow gradients only—no setting, props, textures, complex lighting or decorative elements. All sixteen cells share one continuous gradient plane and ample negative space; never form a literal environment. Hidden visual: across the sixteen background tiles, form one continuous, very faint 45-degree side-profile silhouette derived only from this same subject's identity and hair shape. Split it naturally across the 4×4 grid so the whole canvas reveals the profile, while each individual cell shows only a small abstract light/dark fragment. Use only extremely subtle value changes—no eyes, nose, mouth or detailed facial features. This remains a faceless background motif of the same person, never a second subject. It must not cover the pose figures, weaken their clarity, become a watermark or compete with the grid. Visual hierarchy: first, the sixteen clear pose depictions; second, the quiet minimal gradient; third, the barely visible continuous profile motif. Output only one finished 9:16 vertical image containing the complete 4×4 map. Add no title, explanatory text, logo, watermark, magazine layout text or other lettering. Keep every pose fully clothed, non-explicit and suitable as an editorial pose reference.
더 많은 이미지·동영상 프롬프트 보기 · 이미지 생성 · 동영상 생성 · AI 캔버스에서 만들기