OpenZhenZhen AI 프롬프트 갤러리 · gpt-image-2.5 · image

사진 기하학 해체 이중 레이어 아트 포스터

Photographic Geometry Deconstructed Double-Layer Art Poster

‘사진 기하학 해체 이중 레이어 아트 포스터’은(는) gpt-image-2.5로 만든 작품입니다. OpenZhenZhen에서 원본 프롬프트와 생성 예시를 살펴보고 새로운 창작에 활용해 보세요.

사진 기하학 해체 이중 레이어 아트 포스터 — AI image creation
사진 기하학 해체 이중 레이어 아트 포스터

프롬프트

중국어 프롬프트

将完整的ref-10a-v1实际照片同时作为原始照片和身份参考，并仅通过identity_reference提供这一个图像输入。制作一张独立的高级设计海报，采用3:4竖版构图，不拼接多张照片。画布上下两部分高度严格1:1，各占50%。上半部分保留原始照片，保留同一位成年女性的身份、面部、直视神情、棕色中分长发、头肩姿势、黑色细肩带上衣、结构、真实质感、自然光影与深青蓝色彩氛围。为适配3:2的上方面板，只可在完整且不变形的原图上下自然延展匹配的深青蓝环境背景；不得拉伸、扭曲、裁掉主体或改变主体。仅进行轻微、克制的高级调色，使其具有艺术杂志、独立出版物或展览图像质感。 下半部分提取同一照片中最具识别性的主体、轮廓、结构、姿态与叙事关系，将同一位女性重构为方块化、模块化的几何解构插画。不要完整复制照片，也不要写实描摹；删去无关细节，只保留最能代表她的身份与气质的视觉记忆点，使人一眼看出与上方照片的对应关系。通过几何概括、形状叠压与拼贴式重组重新建立形象。强调方块感、模块感、几何切分与完整大块面。不要把几何拆得过碎或切成大量零散小块；以更少、更大、更完整的矩形、方形、圆角方块、椭圆、三角形和不规则形组织主体。对于这张人像，以方形面部、清晰下颌线和块面化但可辨识的五官进行表现。允许适度夸张或变形，但整体必须成熟、克制、简洁，不可琐碎、幼稚或拼得太散。 采用扁平色块、纸本拼贴构成和现代主义几何语言。让形状像纸片一样重新拼接，形成鲜明但有秩序的剪贴感。删除复杂背景，只保留少量必要辅助形状，不形成第二视觉中心。保持小尺度主体与大面积留白的关系，根据人像自身方向、比例、视觉重心与结构走势安排位置；可偏心、贴边或局部裁切，但须保持女性身份清晰与主次分明。维持疏密节奏、编辑式平衡、正负形、块面轻重、留白呼吸和出版物级版式。 在几何色块上保留细腻纸感、柔和噪点与轻微版画印刷质感，使作品接近纸本拼贴或艺术印刷。避免光滑矢量、3D渲染、照片滤镜和过度精修。从这张实际照片中提取2–4种有辨识度、讨喜的原生颜色，重新整理为高级、耐看的配色系统。本张人像选取实际可见的四种颜色：深青蓝、暖奶油色、深棕黑和柔和珊瑚色。可按需适度提亮、提纯或轻微去灰，但须忠于原图。主色鲜明而不刺眼，辅助色温柔且有层次；整体干净、温暖、时髦、亲和。不要加入原图无依据的色相。避免土黄、脏灰、暗褐主调、荧光色、廉价糖果感和杂乱配色。 可少量加入任意语种的克制小字，例如根据主体身份、情绪、气质或象征意义提炼的简短标题、小字注释、编号或说明片段。文字应小、轻，落在留白或版面边角并与图形自然结合；不要堆叠商业标题，也不要虚构具体地点等事实。整体应由完整大块面的几何解构、克制拼贴、柔和高级配色、纸本颗粒、大面积留白与编辑式秩序共同构成。保持主体身份清晰与模块化艺术语言统一。避免写实描摹、过碎切割、幼稚拼贴、复杂透视、过密细节、卡通感、3D感、模板化海报感和廉价文创感。

영어 프롬프트

Use the complete actual ref-10a-v1 photograph as both the source photo and identity reference, through identity_reference as the sole image input. Create one independent high-end design poster in a 3:4 vertical composition; do not make a multi-photo collage. Divide it into top and bottom areas strictly 1:1 in height, each occupying 50% of the canvas. The top half retains the original photograph. Preserve the same adult woman's identity, face, direct gaze, long center-parted brown hair, head-and-shoulders pose, black thin-strap top, structure, realistic texture, natural lighting and deep teal color atmosphere. To fit the 3:2 top panel, naturally extend only the matching deep-teal environmental background above and below the complete undistorted source image; do not stretch, distort, crop away or change the subject. Apply only slight, restrained high-end color grading for the quality of an art magazine, independent publication or exhibition image. In the bottom half, extract the most recognizable subject, outline, structure, pose and narrative relationship from the same photo. Reconstruct the same woman as a blocky, modular geometric-deconstruction illustration. Do not fully copy the photo or paint realistically. Remove irrelevant detail and retain only the visual memory points that best represent her identity and temperament, so the correspondence with the photo above is immediately recognizable. Use geometric generalization, shape overlap and collage-style reorganization. Emphasize blocks, modules, geometric segmentation and complete large facets. Do not fragment the geometry into many scattered pieces: organize the subject with fewer, larger and more complete rectangles, squares, rounded-corner blocks, ellipses, triangles and irregular shapes. For this portrait, use a square face, clear jawline and blocky but recognizable facial features. Moderate exaggeration or deformation is allowed, but keep the result mature, restrained and concise, never trivial, childish or scattered. Use flat color blocks, paper-collage composition and modernist geometric language. Re-stitch the shapes like paper pieces for a vivid but orderly collage. Delete complex background detail, retaining only a few necessary auxiliary shapes without creating a second visual center. Keep the subject small in scale against large areas of whitespace. Arrange it according to the portrait's direction, proportions, visual focus and structural flow; off-center placement, edge alignment or partial crop are allowed only when the woman's identity and clear hierarchy remain. Maintain rhythmic density, editorial balance, positive and negative shapes, block weight, breathing space and publication-quality layout. Retain fine paper grain, soft noise and a slight printmaking texture over the geometric color blocks, like paper collage or art-print work. Avoid smooth vectors, 3D rendering, photo filters and over-refinement. Extract 2–4 recognizable, pleasing colors present in this actual photo and reorganize them into a high-end, durable system. For this portrait select four that are actually visible: deep teal, warm cream, dark brown-black and soft coral. Brighten, purify or gently desaturate them as needed while staying faithful to the photo. Keep main colors vivid but not harsh, supporting colors gentle and layered; the overall mood is clean, warm, fashionable and approachable. Do not add unsupported hues. Avoid earthy yellow, dirty gray, dark-brown dominance, fluorescent color, cheap candy color and a cluttered palette. Optionally add a small amount of restrained text in any language, such as a short title, tiny annotation, number or brief explanatory fragment, refined from the subject's identity, mood, temperament or symbolic meaning. Keep any text small and light in whitespace or a layout corner, integrated with the graphic; do not stack commercial headlines or invent factual location details. The overall image should unite complete large geometric facets, restrained collage, soft high-end color, paper grain, generous whitespace and editorial order. Preserve clear identity and a unified modular artistic language. Avoid realistic depiction, over-fragmented cuts, naive collage, complex perspective, overly dense detail, cartoonish style, 3D feel, templated-poster appearance and cheap cultural-souvenir aesthetics.

API 모델 보기

더 많은 이미지·동영상 프롬프트 보기 · 이미지 생성 · 동영상 생성 · AI 캔버스에서 만들기