Lumina Mascot LogoLumina
/
Professional 3D Animated Character Portrait | Animator's Brief
人物与角色1:1

Professional 3D Animated Character Portrait | Animator's Brief Using the uploaded reference photo, create a premium-quality 3D animated character portrait of me, inspired by the expressive charm and polished visual storytelling of Pixar-style character design. Character Design: Transform me into a beautifully stylized 3D animated version while preserving my distinctive facial features, face shape, hairstyle, skin tone, and recognizable identity. The character should look instantly recognizable to anyone familiar with the original photo. Maintain natural facial proportions while adding subtle animated charm, expressive eyes, and refined details. Composition: Create a head-and-shoulders bust portrait only, positioned precisely in the center of the canvas. Keep the character relatively small, surrounded by generous, clean negative space on all sides. Use a pure, seamless white background with no additional objects or decorations. Rendering & Lighting: Use high-quality 3D character rendering, smooth yet realistic skin shading, detailed hair, refined facial textures, and soft studio lighting. Keep the skin tone fair and neutral, with accurate natural undertones. Avoid excessive warmth, golden hues, orange casts, or yellowish skin tones. Finishing Details: Give the bottom of the bust a smooth, softly rounded sculptural edge, with a subtle natural shadow directly beneath it to create a gentle floating effect. Maintain clean contours and beautifully polished 3D surfaces. Overall Aesthetic: Premium animated movie character quality, playful yet sophisticated, friendly facial expression, appealing character proportions, cinematic 3D craftsmanship, and professional studio presentation. Technical Requirements: - Aspect ratio: 1:1 (square) - Background: Pure white (#FFFFFF) - Framing: Head and shoulders only - Placement: Precisely centered - Lighting: Soft, balanced studio lighting - Output: High-resolution, crisp and polished - Intended use: WhatsApp profile picture Important: Identity preservation is the highest priority. Do not change my recognizable facial characteristics or invent a different person. The final result should feel like me reimagined as a high-end animated movie character.

GPT Image 2.5
The uploaded photograph is the sole content source. Reimagine the person as a distinctl...
人物与角色1:2

The uploaded photograph is the sole content source. Reimagine the person as a distinctly stylized, cute hand-drawn cartoon portrait on a plain pale mint background. Create a new illustrated character based on the person, with strong facial redesign and moderate likeness. The result must read immediately as a cartoon avatar, with clear simplified shapes and expressive proportions. Actively redesign the head and face rather than preserving photographic geometry. Enlarge the head relative to the visible shoulders, broaden the upper face and cheeks, shorten the distance from the eyes to the chin, and replace the elongated photographic jaw with a compact rounded jaw and small gently tapered chin. Keep the source head tilt, viewing angle, gaze, and expression, but interpret them through this new cartoon construction. For a three-quarter view, use a simplified rounded three-quarter face with appropriate eye overlap and cheek contour, not a realistically modeled portrait. Make the eyes prominent almond-shaped graphic forms with large dark brown or near-black irises occupying most of the visible opening, a small amount of off-white sclera, a confident dark upper eyelid, a few tapered lashes, and one small crisp highlight. Use simple flat iris colors without intricate radial detail. Match the source expression: preserve a wink, narrowed eyes, relaxed eyelids, or an open gaze rather than making every person wide-eyed. Draw eyebrows as distinct curved shapes with a few fine directional strokes. Replace the realistic nose with a tiny curved mark, small nostril indication, or compact peach-colored shape. Omit the modeled nasal bridge, realistic nose projection, detailed nostril anatomy, and eye-socket shading. Shorten and simplify the mouth into a small expressive coral or muted rose shape. Preserve whether the lips are closed, parted, smiling, or pouting, while removing realistic lip texture and volumetric modeling. Use broad warm blush patches on the cheeks, optionally with a few short diagonal pencil marks. Retain only the strongest identity anchors: the approximate face-shape impression after cartoon redesign, skin-tone family, hair color, hairstyle silhouette, parting, expression, age and gender presentation, and distinctive visible accessories. Preserve the source clothing categories, main colors, neckline, headwear, glasses, jewelry, and other existing personal items. Do not preserve exact facial measurements or photographic anatomy. Do not add accessories or replace the subject with a generic character. Draw hair as a few large, clear dark shapes and overlapping curved locks. Add selective brown pencil highlights and a limited number of fine loose strands around the silhouette. Preserve braids, curls, tied hair, bangs, and their placement when present. Keep broad hair masses dominant; avoid covering the entire hairstyle with thousands of fine realistic strands. Use slightly grainy dark brown pencil contours with natural variation in pressure, combined with broad matte gouache-like color fills. Keep skin mostly flat and warm, with only a few soft peach shadows beneath hair overlaps and the chin. Apply subtle paper grain across the illustration without making the texture dominate. Simplify clothing into broad shapes with sparse sketch lines and two or three muted shadow regions. The overall finish should be clean, soft, colorful, and visibly handmade. Replace all scenery with a uniform pale mint or pastel aqua background. Preserve the source crop and visible body range, including head angle and shoulder direction. Adapt the enlarged head within the existing framing without inventing unseen body parts. Use a close portrait composition with the face and hair as the main focus. Prioritize the redesigned cartoon face over precise likeness: rounded cheeks, shortened lower face, prominent dark eyes, tiny symbolic nose, small expressive mouth, organized hair masses, and restrained pencil texture. Avoid realistic colored-pencil portraiture, traced facial proportions, long anatomically modeled noses, deep eye sockets, skin pores, dense cross-hatching, realistic strand-by-strand hair, glossy rendering, 3D effects, extreme infant proportions, and elaborate backgrounds. Output one completed cartoon portrait only, with no original photograph, comparison layout, padding, text, interface elements, signature, or watermark.

GPT Image 2.5
A close-up, high-contrast black and white portrait of a m...
人物与角色3:4

A close-up, high-contrast black and white portrait of a man looking directly into the camera with an intense expression. He has thick, voluminous messy dark hair, a neatly trimmed beard, and a mustache. He is wearing dark aviator-style sunglasses with a double bridge bar, which he lowers slightly down his nose with his right hand, peering directly over the top rims. On the ring finger of his right hand, he wears a prominent square signet ring with a dark center inlay. He is dressed in a dark button-up shirt, left open at the collar to reveal a hint of his chest. The background is a dark, out-of-focus brick wall with faint horizontal lines. The lighting is soft and directional, coming from the front-top oblique angle, highlighting his forehead, nose bridge, and knuckles while leaving the rest of the scene in deep shadow.

Banana Pro
设计为上下双拼风格,3:4竖版构图。
人物与角色3:4

设计为上下双拼风格,3:4竖版构图。 上半部分展示用户上传的原图主体,重点提取人物、猫、狗等动物主体,尽量保留主体真实外貌特征、发型/毛发轮廓、脸型、五官、姿态、服饰或毛色配色与整体气质,环境只作为弱化背景处理,建筑风格不作为重点。 下半部分将同一主体转译为ASCII 字符构成的极简平面海报风格,整体采用 ons design 极简平面设计语言,使用 ASCII 二进制 0 和 1 数字字符拼接出完整主体形象;如果上半部分是人物,则由字符拼接出完整人物主体;如果上半部分是宠物,则由字符拼接出完整宠物主体。 色彩上延续克制审美:浅粉色数字字符用于主体柔和区域或花瓣般轻盈部分,淡绿色字符用于辅助轮廓、服饰局部、毛发层次或次要结构,黑色字符用于主体结构重心、底部承托形体、阴影结构或局部几何稳定元素。 背景为浅米白色细网格背景,画面散落低饱和莫兰迪色块作为装饰:裸粉、浅灰、浅豆绿方形装饰块。整体大面积留白,气质清冷、素雅,带有新中式极简氛围。 保持哑光平涂质感、柔和光影、无强烈明暗阴影,字符像素细节精致,整体具有高清印刷海报质感。 去除所有文字内容,包括但不限于标题、双语说明、字段名、落款、年份、品牌名等,不生成任何可读文字,只保留信息海报般的编排感与字符构成视觉。 无水印、无多余杂物、无额外人物。

GPT Image 2.5
Create image
人物与角色9:16

Create image [BRAND NAME]: {fill} [CAMPAIGN / SERIES]: {fill} [MAIN TITLE]: {fill} [TAGLINE]: {fill} [SPATIAL TYPE SYSTEM]: giant 3D letters / floor typography / letter-shaped openings / folded-wall typography [PRIMARY COLOR]: {fill} [SECONDARY COLOR]: {fill} [NUMBER OF PEOPLE]: {fill 6-8 people} [ASPECT RATIO]: 9:16 Create a high-end editorial campaign poster where typography becomes a real architectural space instead of a flat graphic placed behind people. Build the entire composition around one clear perspective system. Turn the main title into a physical spatial structure: oversized letters people can walk around, step across, pass through, disappear behind, or interact with through real depth and occlusion. Arrange multiple adult figures at different distances within the same perspective field. Their scale must change naturally with depth. Avoid equal spacing, grids, lineups, or team-photo compositions. Some figures should be large in the foreground, some partially hidden by the typography, and others much smaller in the distance. Make people and type physically interlock: bodies crossing letter edges, figures appearing through negative spaces, typography blocking parts of the body, and architecture creating multiple foreground, midground, and background layers. Keep the environment minimal so the main spatial idea stays immediately readable. Use bold geometric forms, large clean color fields, strong directional daylight, and realistic architectural shadows. Use realistic fashion photography for the people: natural skin texture, believable hair, accurate anatomy, real fabric folds, varied poses and subtle movement. Avoid CGI skin, duplicated people, floating feet, or overly staged model poses. Typography should remain readable while following real perspective, scale, surface angle, and depth. Supporting text should be minimal, clean, and secondary to the main visual idea. The final result should feel like a real 2026 fashion, lifestyle, or cultural campaign photographed inside a purpose-built typographic installation — bold at thumbnail size, visually simple at first glance, but full of spatial detail on closer inspection. Avoid flat poster typography, random text, generic collage layouts, PPT-style grids, excessive decorative elements, unnecessary logos, numbering, watermarks, sparkle symbols, and cheap 3D effects.

GPT Image 2.5
A rugged, charismatic middle-aged man with long wavy swept-back hair, full beard and mu...
人物与角色2:3

A rugged, charismatic middle-aged man with long wavy swept-back hair, full beard and mustache, wearing vintage aviator sunglasses, a black ribbed turtleneck and tailored dark blazer. Dramatic close-up three-quarter portrait, looking slightly upward with a confident, contemplative expression. Highly detailed vintage comic-book / graphic-novel illustration, intricate hand-inked linework, dense cross-hatching, stippling and engraved textures, bold black shadows, limited palette of cream, charcoal, burgundy, muted teal and warm amber. Retro 1970s poster aesthetic, halftone print texture, screen-print effect, strong rim lighting, textured turquoise background fading into deep black, cinematic composition, sharp facial details, expressive hair strands, sophisticated editorial artwork, high contrast, vertical poster, 4:5 aspect ratio.

GPT Image 2
GPT Image 2 on ChatGPT
人物与角色2:3

GPT Image 2 on ChatGPT\n\nPrompt:\n\nCreate a premium 2×2 editorial portrait collage featuring four globally recognized personalities.\n\nTop Left: Elon Musk\nTop Right: Sam Altman\nBottom Left: Cristiano Ronaldo\nBottom Right: Donald Trump\n\nEach panel features a dramatic black-and-white photorealistic portrait with a subtle geometric digital painting style. The subjects are looking slightly away from the camera with confident expressions. Clean white background with black ink splashes and modern editorial design elements.\n\nDisplay large bold uppercase names vertically on the left side of each panel:\nELON MUSK\nSAM ALTMAN\nCRISTIANO RONALDO\nDONALD TRUMP\n\nBelow each name, add a short professional title:\nFounder of SpaceX, Tesla & X\nCEO of OpenAI\nFootball Legend\n45th President of the United States\n\nMinimal monochrome color palette, luxury magazine cover aesthetic, high contrast lighting, ultra-sharp facial details, premium typography, balanced composition, cinematic shadows, realistic skin texture, editorial poster design, modern branding, 8K HDR, masterpiece quality.\n\nNegative Prompt:\nlow quality, blurry, bad anatomy, distorted face, duplicate, extra limbs, watermark, logo, cropped, oversaturated, noisy background, misspelled text, artifacts.

GPT Image 2
将上传的每一张照片转换成独立的高级编辑比照海报。每张照片都要独立地处理,避免将图片或主题、颜色或不同照片的元素结合。使用严格的3:4比例肖像画布,水平分成精确的两种平等的50...
人物与角色3:4

将上传的每一张照片转换成独立的高级编辑比照海报。每张照片都要独立地处理,避免将图片或主题、颜色或不同照片的元素结合。使用严格的3:4比例肖像画布,水平分成精确的两种平等的50%部分,中间部分保持无缝连接。上半部分 — 原创摄影:忠实地保留原创照片,包括主题身份、面部特征、表情、姿势、比例、构图、相机角度、框架、真实的纹理、自然的照明和原创的色彩氛围。只添加微妙的专业编辑色彩评分,用于精致的艺术杂志或展览美学。如果必要,自然地扩展周围的背景,不要伸展、扭曲、替换或修改任何主题。保持最接近的原创照片画法。不要添加文字、边框、图形或装饰元素。下半部分 — 表达性的中国墨水涂鸦画:重新解释同一张照片作为真实的传统中国手工墨水画(Xieyi / Sumi-e)在稍微变老的温暖象牙色Xuan米纸上,从上方直接观察。从照片中抽取最可识别的主题、轮廓、手势、空间关系和构图结构,保留它们的相对位置、比例和视觉层次。重建场景使用表达性的书法笔触、五到六种色调的黑色和灰色墨水、干刷纹理、飞白笔触、湿墨涂鸦和宽广的空白空间。避免摄影真实主义、精确的轮廓、过度的细节和字面追踪。遵循“一幅画”的原则,宽广的空白空间,一亿点的注重。画面应该占占据下方画板的约80%,留出自然的空白米纸边缘。 用淡淡的墨水和大气中的空虚感呈现远处元素,中间地平的主题用层次分明的中间地平的主题,用多层的中间地平的墨水和深色墨水涂抹,前景用大胆的笔触涂抹。保留原图的左右和上下排列。只在现有有意义的视觉元素中,例如服装、雨伞、灯笼、花朵、屋顶细节或焦点物体中,引入只一个小巧精心地位置的紫色点缀。紫色点缀必须极其小、被限制且自然地融入画作中,从而永远不出现作为单独的图形、标志、印章或装饰性标记。绝对无封印和无框规则:不要在整个图片中生成任何传统的印章、艺术家的签名印章、红色方块、红色长方形、红色轮廓、红色边缘、角落标志、盒子标志、字母标志、装饰性标志、水印、标志或印章等。下方右角必须完全不受封印、标志、红色标记、标记盒子和装饰性元素的影响。所有外边和纸张边缘必须保持干净、自然质地和未装饰。颜色限制:只在画作中使用五到六种单色黑色和灰色墨水,温和地呈现出黄色米纸,使用温和地黄色米纸和一条细小的紫色点缀,引入真实的墨水吸收。无需使用数字渐变、人工纹理、真实的阴影、或数字过滤的摄影。字体:在画作下方只在字母上精心绘制的英文字母中添加一个极小的线条,源自照片的可识别位置或背景。 如果精确的地点不明,请使用简体中文的简短描述英文名称,而不是编造地理位置。除本文、字母、数字、标志、字幕、标签、签名、水印或印刷品在海报上任何地方,除本文中已可见的真实细节外,其他文本、字母、数字、标志、字幕、标签、签名、水印或印刷品在海报上都没有。最终的美学:真实摄影 × 经典的中国式表达墨画 × 简约的编辑设计。安静、诗意、克制、氛围感、复杂且空间丰富。创建一个令人惊叹的对比,照片真实与表达墨画抽象之间保持清晰的视觉对照。严格避免使用色彩鲜艳的插图、真实的水彩风景、过度的紫色、中国式的印章、红色角印章、红色方形轮廓、装饰框架、卡通美学、商业广告布局、重复的模板、发明的主题和不必要的细节。海报必须包含绝对没有红色印章、红色框架、红色盒子或底右角的印章或印章。

GPT Image 2.5
Use the two provided reference images as the ONLY visual references for this creation.
人物与角色2:3

Use the two provided reference images as the ONLY visual references for this creation. RECREATE THE SAME TWO PORTRAITS shown in the reference images. Do not redesign, reinterpret, or change the subjects. REFERENCE 1 — MALE: Preserve the exact same male subject identity, facial structure, face shape, eyes, eyebrows, nose, lips, jawline, skin tone, dark wavy hair, facial proportions, cowboy hat, clothing, pose, expression, camera angle, lighting, and overall appearance from the reference image. REFERENCE 2 — FEMALE: Preserve the exact same female subject identity, facial structure, face shape, eyes, eyebrows, nose, lips, jawline, skin tone, long dark wavy hair, facial proportions, cowboy hat, clothing, pose, expression, camera angle, lighting, and overall appearance from the reference image. COMPOSITION: Recreate the exact same three-panel vertical editorial composition shown in the references. Use three tall vertical portrait panels with clean white gaps between them. The center panel must be taller than the left and right panels. The subject's face, hat, hair, shoulders, and clothing must continue naturally across the three panels exactly like the reference. STYLE: High-end luxury fashion editorial photography, photorealistic, warm burnt-orange monochromatic color palette, solid orange studio background, soft cinematic studio lighting, realistic skin texture, detailed individual hair strands, realistic fabric texture, premium fashion magazine aesthetic, sharp facial details, natural proportions, professional photography. CRITICAL IDENTITY LOCK: Do not change the male or female identity. Do not generate new faces. Do not substitute different people. Do not alter facial proportions. Do not change hairstyle, hat, clothing, expression, pose, or camera angle. Keep both subjects visually identical to their respective reference images. CRITICAL VISUAL MATCH: Match the reference images as closely as possible in composition, framing, proportions, colors, lighting, shadows, wardrobe, background, panel placement, facial positioning, and overall photographic appearance. No text, no logos, no watermark, no additional people, no extra objects, no face distortion, no duplicated facial features, no anatomy errors.

GPT Image 2.5
Create a 1970s black-and-white close-up portrait in a win...
人物与角色2:3

Create a 1970s black-and-white close-up portrait in a windswept outdoor setting, black and white noise.Camera/look: eye-level, front ¾ view (~30° off-axis); 85–105 mm portrait perspective; vertical 2:3, tight head-and-shoulders crop; shallow DoF f/2.8–f/4; subject centered.Lighting: hard natural sunlight from upper camera-left with minimal fill; crisp shadows; subtle rim on hair; sky as bright mid-tone background; no blown highlights.Pose & expression: subject faces camera, chin slightly down, lips parted with a cigarette held between them; gaze relaxed/half-lidded; hair blown across the face by wind, partially covering one eye.Wardrobe & styling: dark collared shirt with open neck; thin black neck scarf tied loosely; no logos/jewelry.Grade: high-contrast monochrome with rich blacks, bright highlights, medium–heavy film grain (Tri-X/HP5 vibe), slight halation, faint vignette; keep skin texture natural (no plastic smoothing).Composition: simple sky background, no extra props; clean verticals; ensure sharp focus on near eye and cigarette.Output: high-resolution, editorial/classic cinema mood; no text or watermarks.

Banana Pro
Photo × Line Art | Real-Scene Extension Travel Poster
人物与角色2:3

Photo × Line Art | Real-Scene Extension Travel Poster Transform the uploaded photo into a “Photo × Line Art | Real-Scene Extension Travel Poster.” The canvas must use a vertical 2:3 aspect ratio, divided horizontally into two exactly equal 50/50 sections, with the dividing line positioned precisely at the center. TOP 50% — ORIGINAL PHOTO Use the original uploaded photograph directly in the upper half. Preserve the original: * subjects and people * buildings, animals, and objects * lighting and colors * composition * photographic detail and sharpness Do not redraw, blur, stretch, mirror, duplicate, or arbitrarily alter the photograph. Only minimal proportional cropping or natural environmental extension is allowed when necessary to fit the required format. CONNECTION — PHOTO → LINE ART Analyze the original photograph and identify one real visual element that is best suited for extension, such as: water, a road, rope, cable, beam of light, bridge, pier, tree branch, mountain slope, architectural contour, mooring post, reflection, or another naturally occurring element. This element must begin from its exact position in the original photograph, cross the center dividing line continuously, and gradually transition from realistic photographic texture into simple black line art. The extended element must have a clear visual interaction with the small illustrated character in the lower half. There must be no gap, misalignment, restart, or disconnected transition at the center line. The illustrated character must not feel unrelated to or separate from the photograph. BOTTOM 50% — MINIMAL LINE ART Use a clean, neutral warm-white / ivory paper background with an extremely subtle fine-fiber texture. Add one small black pen-line character, occupying approximately 7–10% of the total poster height. Create an original action based on the content of the photograph. For example, the character may be: pulling, holding, folding, pushing, climbing, observing, walking, resting, or interacting with the extended photographic element in another meaningful way. The character should have: * no clearly defined facial features * highly simplified clothing and props * thin black pen lines * slightly irregular, handmade strokes * relaxed and natural linework HANDWRITTEN PHRASE Based on the mood and content of the photograph, generate a gentle 2–5 word English phrase. Render it in small handwritten lettering. The phrase should function as a subtle emotional note or poetic caption — not as a main headline. OVERALL STYLE The final result should feel: premium, restrained, poetic, editorial, minimal, and conceptually clever. Use the transition between photography and line art as a subtle visual metaphor. Maintain generous negative space. For every new uploaded photograph, create a different connection concept based specifically on that image. DO NOT Do not create: * completely disconnected photo and illustration sections * floating characters * isolated icons * oversized characters * detailed facial features * chibi or anime characters * thick/heavy outlines * dense sketching * a full line-art recreation of the entire scene * infographic-style compositions * collage grids * postage-stamp elements * decorative borders * watermarks * phone UI elements * overly yellow paper * oversized typography

GPT Image 2.5
创作一张受上传参考图启发的优质编辑风格插画旅行海报...
人物与角色2:3

创作一张受上传参考图启发的优质编辑风格插画旅行海报。保持整体构图、人物造型、摩托车位置、排版布局和调色板的视觉一致性,但使用全新的措辞和表达。\n\n一位时尚的年轻东亚女性自信地站立在一辆经典复古赛车摩托车旁边。她留着一头略显蓬松的深棕色短波波头,带有松散的刘海,蓝色头戴式耳机挂在脖子上。她穿着一件蓝白复古赛车夹克,上面布满彩色赛车补丁,内搭黑色上衣,下穿深海军蓝格纹百褶迷你裙,白色中筒袜和厚底白色运动鞋。一个黑色单肩包挂在一侧肩膀上。\n\n摩托车醒目地置于构图的左下方,采用经典的白、蓝、黄、红赛车涂装,细节丰富的整流罩、车轮、后视镜、机械部件,以及一个配套的赛车头盔放在摩托车上。\n\n使用柔和的手绘编辑插画风格,具有写实解剖结构和细致纹理,微妙的水彩晕染,铅笔般的线条,柔和的颜料飞溅,以及精致的复古时尚杂志美学。背景应为暖米白色纹理纸,留有充足的干净留白。\n\n在左上方,添加大号优雅手写毛笔字“Tokyo”。在其下方,放置小号大写字母间距排版文字:\n\nSAME\nROADS\nDIFFERENT\nDREAMS\n\n在右侧,包含三个不规则的小型手绘色块,分别为深海军蓝、金黄色和柔和砖红色,随后是一个浅米色色块。\n\n在右下方,添加极简编辑排版文字:\n\n2026\n—\nGOOD\nTHINGS\nAHEAD\n\n平衡的时尚海报构图,精致的日式街头赛车氛围,高级杂志设计,微妙的纸张纹理,克制的复古调色板,艺术感强但细节丰富的人物和摩托车,整洁的排版布局,优雅的留白,垂直 3:4 宽高比。

GPT Image 2.5
Masterpiece, ultra-detailed hyperrealistic anime realism, full-body portrait of an extr...
人物与角色2:3

Masterpiece, ultra-detailed hyperrealistic anime realism, full-body portrait of an extraordinarily beautiful young Korean female shaman, Korean Mudang (무당), elegant and mysterious, realistic East Asian facial features, luminous natural skin, delicate facial structure, calm commanding gaze, subtle cinematic beauty.\n\nShe wears an elaborate Korean shamanic ceremonial costume inspired by traditional hanbok and Muism: enormous flowing dark navy outer robes, voluminous white inner garments, deep crimson lining, intricate antique-gold embroidery, layered ceremonial belts, Korean ornamental pendants, ritual jewelry and traditional hair ornaments. Long black hair arranged in an elaborate high ceremonial hairstyle with delicate dangling accessories.\n\nOne hand rests naturally on an ornate traditional Korean sword in its decorated black scabbard; the other hand is gracefully raised, fingers elegantly extended, holding multiple aged Korean shamanic talismans, ritual paper charms, beads and red-blue tassels.\n\nThe garments are extremely voluminous and monumental, with long sleeves and layered fabric sweeping dramatically across the composition, realistic silk texture, natural folds, subtle fabric weight and exquisite embroidery.\n\nBehind her is a huge expressive black ink-brush circle painted across a clean warm-white background, bold Korean ink-painting energy, scattered ink splashes and dry-brush textures, creating a powerful minimalist contrast with the highly detailed character.\n\nStrong Korean Muism visual identity, sacred ritual atmosphere, mysterious but elegant, ancient spirituality fused with contemporary fantasy character design.\n\nHyperrealistic anime aesthetic, realistic skin pores, individual hair strands, physically believable silk and metal materials, cinematic soft lighting, subtle shadows, extremely refined costume details, premium Korean fantasy film character-design quality, sophisticated editorial composition, clean background, no environmental scenery, no Tibetan elements, no Chinese Taoist aesthetic, no Japanese shrine elements.\n–ar 9:16

GPT Image 2
Create a premium Pinterest editorial contact sheet on a pure white background.
人物与角色1:1

Create a premium Pinterest editorial contact sheet on a pure white background.\n\nThe page contains exactly 16 equally sized rectangular panels arranged in a perfect 4×4 grid with thin black borders and consistent spacing.\n\nAlternate every panel between:\n\n1. an ultra-realistic portrait photograph\n\n2. a clean black-and-white manga-style line drawing of the SAME person.\n\nEvery sketch must correspond to the adjacent photograph, preserving the same hairstyle, facial structure, pose, and expression.\n\nUse different people across the sheet (both men and women), with varied hairstyles and poses. Some face forward, some in 3/4 view, some side profile.\n\nThe line drawings must look like professional character design studies using clean black ink, minimal shading, expressive eyes, and confident line work.\n\nThe photographs should feel like editorial fashion portraits with soft natural lighting.\n\nMinimalist magazine layout, Pinterest aesthetic, concept art presentation, white background only, no text, no watermark, no logos, perfectly aligned grid, high-end art book quality.

GPT Image 2
A full-body studio portrait of a handsome man with a neatly trimmed, full
人物与角色2:3

A full-body studio portrait of a handsome man with a neatly trimmed, full beard and dark messy hair, sitting on a tall black wooden barstool against a seamless, solid light grey background. He is dressed in a sophisticated all-black ensemble, featuring a tailored slim-fit suit jacket worn unbuttoned over a snug knitted turtleneck sweater, matching black trousers, and highly polished black leather oxford dress shoes. He is wearing pair of glossy black-rimmed round eyeglasses with smoky, semi-translucent lenses that slightly obscure his eyes while looking directly forward with a confident expression, his left knee angled outward with his left hand resting on his thigh, while his right foot rests flat on the studio floor. The lighting is soft and diffused from a studio softbox source, casting a smooth shadow on the floor extending behind the stool and creating fine, bright highlights on his face and the glossy surfaces of his leather shoes.

Nano Banana Pro
Elegant watercolor fashion illustration of a stylish youn...
人物与角色9:16

Elegant watercolor fashion illustration of a stylish young woman exploring scenic European cities and cozy cafés, soft hand-painted watercolor textures, expressive ink sketch outlines, long wavy dark hair, chic oversized jackets, denim jeans, tote bags, sneakers, vintage camera, candid travel moments, historic streets, warm café lighting, pastel neutral tones, airy white background, delicate brush strokes, luxury fashion editorial meets travel sketchbook diary, Pinterest aesthetic, cinematic sunlight, modern feminine vibe, ultra detailed, elegant urban lifestyle artwork, timeless chic atmosphere, 8K masterpiece.

GPT Image 2.5
Cinematic portrait of a handsome young man with short sty...
人物与角色3:4

Cinematic portrait of a handsome young man with short styled hair and a trimmed beard, wearing thin metal aviator glasses, softly illuminated by a glowing golden circular halo behind his head. Warm amber lighting, dramatic low-key background, shallow depth of field, hyper-realistic skin texture, subtle reflections in the glasses, moody and futuristic aesthetic, studio photography, 85mm lens look, ultra-detailed, high contrast, dark minimal background, symmetrical composition, elegant and modern tone.

Banana Pro
A stylish South Asian man stands in a newspaper-lined int...
人物与角色3:4

A stylish South Asian man stands in a newspaper-lined interior, adjusting his suspenders with one hand while touching his neatly sculpted beard with the other, projecting a composed editorial-fashion presence. Soft daylight enters from the upper right, creating bright diffused highlights across his face and white checked shirt while preserving gentle shadows beneath his jaw, collar, and arms. His round tinted glasses produce subtle specular reflections, while the leather suspenders, metal wristwatch, textured belt, and finely pleated dark trousers provide contrasting tactile surfaces. He occupies the center-right foreground, sharply defined against a shallow-depth-of-field background of overlapping newspaper pages, faded typography, and muted beige-gray tones, with a slight vignette framing the portrait. * **Style:** Editorial fashion photography, softly retouched photorealism * **Key elements:** South Asian man, suspenders, newspaper wall, checked shirt, wristwatch * **Lighting:** Soft diffused daylight from upper right, neutral-warm color temperature * **Camera:** Medium portrait shot, 50mm lens, eye-level angle

Banana Pro
Use the uploaded photograph as the sole visual and content reference.
人物与角色1:2

Use the uploaded photograph as the sole visual and content reference. Transform the photograph into one standalone monochrome editorial ink illustration printed on warm off-white, lightly textured paper. Output only the finished illustrated page—not the original photograph, not a before-and-after comparison, not a split-screen layout, and not a collage. Preserve the source image’s main subject, recognizable silhouette, viewpoint, pose, expression, gesture, proportions, orientation, important contours, and essential interaction with objects. The drawing must remain immediately recognizable as the same subject and moment. This transformation must work for people, animals, food, everyday objects, vehicles, and still-life scenes. Identify the main subject automatically and give it visual priority. Preserve distinctive features that carry recognition—for example facial expression, hairstyle, eye direction, open mouth, animal markings, ear shape, cup handle, latte pattern, bicycle frame, basket, flowers, spokes, or other defining structures visible in the source. Redraw the subject using expressive black dry-ink, rough charcoal, grease-pencil, and broken lithographic marks. Use energetic, imperfect contours; scratchy repeated strokes; irregular pressure; dry-brush gaps; grainy black deposits; loose cross-hatching; scattered speckles; and partially missing edges. Let some lines overlap or overshoot naturally. The result should feel physically drawn by hand and mechanically printed on absorbent paper. Use only deep black ink and the warm ivory color of the paper. Build form through line density, broken black masses, negative space, sparse hatching, and raw directional strokes. Do not use continuous grayscale shading, smooth gradients, photographic tonal modeling, or clean digital fills. Simplify secondary details aggressively. Remove the photographic background unless a background element is necessary to identify the subject or explain its pose. Reduce necessary context to a few loose grounding lines, faint shadows, fragmented texture, or minimal environmental marks. Keep large areas of clean paper visible. For human portraits, preserve the approximate face shape, hairstyle, gaze direction, expression, mouth position, shoulders, and visible clothing category. Draw the eyes, nose, and mouth with sparse but expressive marks. Emphasize the rhythm and volume of the hair with dense, tangled strokes when appropriate. Do not beautify, change identity, or replace the expression. For animals, preserve species, head shape, beak or muzzle, eye placement, markings, posture, and expression. Use denser black marks in dark feathers or fur and paper-white reserve areas in lighter regions. For objects, preserve functional geometry and recognizable construction while allowing loose perspective and handmade irregularity. Keep important circular, mechanical, or structural relationships readable without turning the image into a precise technical drawing. Create a balanced editorial-page composition. Place the main illustration in the central or right portion of the page, occupying most of the visual area, while reserving a clean vertical block of negative space on the left for optional typography. If the original subject requires the full width, reduce the text area instead of distorting or cropping the subject. Optional typography module: Only when the user supplies exact wording, place a restrained editorial text block in the negative space using elegant high-contrast serif typography. Arrange it in this hierarchy: [ISSUE NUMBER] [SHORT TITLE] — [ONE-LINE CAPTION] Use black ink, left alignment, generous line spacing, and clear typographic hierarchy. The title may wrap naturally across two short lines. Keep the text visually secondary to the illustration. Use only the exact text supplied by the user. Do not invent issue numbers, titles, captions, names, dates, slogans, or quotations. If no text is supplied, omit typography and use the empty area as intentional negative space. Overall aesthetic: raw editorial spot illustration, handmade black-ink sketch, vintage literary magazine page, dry lithographic print, expressive observational drawing, tactile paper grain, restrained monochrome design, spontaneous but carefully composed. Avoid: displaying the source photograph, top-and-bottom comparison, split screen, photo-and-drawing collage, realistic pencil portrait, smooth graphite gradients, gray wash, watercolor, colored ink, clean vector outlines, uniform line weight, polished digital illustration, comic-book inking, anime, manga, photorealism, precise architectural drafting, excessive background detail, solid black background, decorative borders, changed pose, changed expression, generic replacement subject, missing defining features, invented objects, malformed anatomy, extra limbs, gibberish text, logos, signatures, or watermarks.

GPT Image 2.5
Transform the attached image into a minimal navy-and-white hand-drawn die-cut sticker i...
人物与角色2:3

Transform the attached image into a minimal navy-and-white hand-drawn die-cut sticker illustration.\n\nTreat the uploaded image as the sole source of content. First identify the primary subject or subjects. Support a single person, a single pet, two people, two pets, or one person with one pet. Preserve the exact number and type of primary subjects. Preserve their left-to-right order, relative scale, interaction, pose, gesture, head angle, gaze direction, expression category, clothing categories, worn accessories, and any object that is essential to the action or story. Preserve the source crop and viewpoint; if the source is a close-up, keep it a close-up. Never invent off-frame bodies, limbs, clothing, or scenery.\n\nCreate a recognizable but deliberately simplified character interpretation, not a traced portrait. Preserve identity through a small set of distinctive cues: face shape, hairstyle and silhouette, key accessories, characteristic expression, posture, and relationship between subjects. Do not reproduce pores, wrinkles, individual eyelashes, realistic eye anatomy, or photographic skin detail. For adults, use mild cartoon simplification only: slightly enlarged heads, moderately shortened torsos, and simplified limbs while clearly remaining adult. Avoid baby proportions, swollen cheeks, toddler bodies, oversized heads, or extreme chibi anatomy.\n\nUse clean, friendly editorial doodle art with rounded, moderately thick, nearly uniform strokes and only slight hand-drawn variation. Outer contours may be slightly stronger than interior lines. Faces must be extremely economical: dot or short curved eyes, short brows, a one-stroke nose, and a simple mouth that retains the original expression. Render all outlines and dark fills in one deep navy blue, approximately #173B7A; never use pure black. Use white as the default fill. Convert clearly dark source areas such as dark hair or dark clothing into solid navy shapes, using only minimal white highlight cutouts when needed for readability. Keep human skin white and unshaded.\n\nThe default palette is strictly deep navy, white, and a very pale neutral-gray shadow. Do not introduce a third color unless the source contains at least two separate, clearly purple or lavender accessories or travel props that are narratively important; only then may those corresponding items use one flat muted lavender accent. Otherwise use no lavender and no additional color.\n\nRemove the entire original environment. Remove architecture, landscape, ordinary furniture, floor texture, incidental objects, unrelated people, signage, interface marks, logos, captions, and watermarks. Place the simplified subject group alone on a clean, opaque white square canvas with generous margins. Wrap the complete group in one continuous wide white die-cut sticker border. Add a very subtle soft pale-gray shadow offset slightly down and right so the white border remains visible against the white canvas. No transparency.\n\nKeep the original grouping and emotional relationship. Compress only empty space; do not change the narrative. Use a balanced centered composition, but keep asymmetry when it comes from the source. Include only the body parts visible in the source and retain the source framing. The finished image should feel like one cohesive sticker, not multiple separate stickers.\n\nAvoid photorealism, line-art tracing, realistic portrait rendering, detailed anatomy, pencil or charcoal texture, watercolor, oil paint, anime or manga rendering, extreme chibi style, 3D rendering, gradients, glow, paper texture, grain, halftone, collage, complex lighting, detailed backgrounds, added props, altered subject count, duplicated limbs, malformed hands, text, captions, logos, signatures, and watermarks.\n\nReturn exactly one finished square sticker illustration.

GPT Image 2.5
Create a premium cinematic poster-style portrait of the uploaded person in a clean beig...
人物与角色2:3

Create a premium cinematic poster-style portrait of the uploaded person in a clean beige studio environment, aspect ratio 4:5. The composition closely follows the reference design. A large side-profile portrait of the uploaded person dominates the background, occupying approximately 70-75% of the frame height, facing left with a calm, confident expression. Behind the large portrait, add strong horizontal motion blur streaks extending toward the right side, creating a dynamic sense of movement. In the foreground, place a smaller full-body version of the uploaded person standing in front of the larger portrait, occupying approximately 30-35% of the frame height. He is wearing stylish beige-tan casual clothing matching the reference aesthetic: a relaxed overshirt/jacket, light-wash loose-fit jeans, and casual sneakers. Add modern transparent eyeglasses. Both hands are casually tucked into his trouser pockets. His stance matches the reference exactly: one leg crossed naturally in front of the other, leaning slightly with a relaxed, confident posture. He is looking slightly toward the brand text area rather than directly at the camera. Place a large modern branding text element behind the foreground figure, partially obscured by his body, using the word \"PROMPT\" in bold futuristic typography. The text should feel integrated into the design, with subtle transparency and layered depth effects. Add minimal geometric UI-style design elements and soft translucent panels around the text area for a contemporary creative-director look. Soft editorial lighting, luxury fashion campaign aesthetic, high-end digital poster design, ultra-detailed facial features, sharp focus on the subject, subtle shadows beneath the standing figure, clean background, premium color grading, realistic textures, professional photography, modern creative branding artwork. No logos, no watermarks, no extra text except \"PROMPT\".

Nano Banana Pro
一位自信的年轻男士的肖像,身穿量身定制的...
人物与角色2:3

一位自信的年轻男士的肖像,身穿量身定制的黑色西装、白色正装衬衫和修身黑色领带,双手随意插在口袋里,一条腿交叉站立,穿着锃亮的黑色牛津皮鞋,系着整洁的皮带;发型干净利落,胡须修剪整齐,温暖的中等肤色,沉着严肃的表情,自然的皮肤渲染和细致的修饰。工作室场景,柔和的暖米色背景,背景中融入了一幅该男士戴着墨镜的大幅柔焦单色肖像,如同双重曝光。左上方柔和的电影主光,柔和的轮廓光将主体分离,西装面料上柔和的阴影和精致的高光,浅景深,中长焦构图,高细节时尚摄影,优雅,现代,略带忧郁的色彩分级,微妙的渐晕,清晰的纹理,真实的材质和精致的构图。

GPT Image 2
Transform the uploaded photo into a beautiful hand-painted storybook illustration with...
人物与角色2:3

Transform the uploaded photo into a beautiful hand-painted storybook illustration with a charming vintage children's book aesthetic. Art Style: Traditional gouache and watercolor painting, visible paper grain, soft brushstrokes, subtle canvas texture, warm earthy colors, delicate hand-drawn outlines, and rich painterly details. Character: Preserve the original subject's identity, facial features, hairstyle, clothing, pose, and distinctive details while translating everything into a charming illustrated form. Use expressive eyes, soft facial features, and naturally flowing shapes. Environment: Recreate the original surroundings as a beautifully illustrated storybook setting, adding intricate environmental details, organic textures, and a dreamy atmosphere. Color Palette: Warm autumn oranges, golden yellows, burnt reds, earthy browns, muted beige, and soft sky blues. Maintain rich but harmonious colors. Lighting & Mood: Soft natural daylight, warm nostalgic atmosphere, whimsical storytelling feeling, cozy and magical visual appeal. Final Look: A richly detailed, full-frame traditional children's storybook illustration, resembling a hand-painted vintage picture book. Absolutely no photorealism, 3D rendering, digital glossy effects, or artificial AI appearance.

GPT Image 2.5
{
人物与角色2:3

{ "metadata": { "image_type": "photograph", "primary_purpose": "editorial" }, "composition": { "rule_applied": "rule of thirds", "aspect_ratio": "2:3", "layout": "layered", "focal_points": [ "Subject's face and sunglasses", "Flowing silhouette of the black overcoat" ], "visual_hierarchy": "Eye starts at the subject's face/sunglasses, travels down the sweeping lines of the black coat to the stride of the grey trousers, then moves back to register the black vehicle and architectural background", "balance": "asymmetric" }, "color_profile": { "dominant_colors": [ { "color": "Deep Black", "hex": "#111111", "percentage": "45%", "role": "primary subject clothing and vehicle" }, { "color": "Slate Grey", "hex": "#5a5d60", "percentage": "20%", "role": "trousers and street surface" }, { "color": "Navy Blue", "hex": "#1c2331", "percentage": "25%", "role": "background building facade" }, { "color": "Warm Amber", "hex": "#ecae62", "percentage": "2%", "role": "accent light from sconce" } ], "color_palette": "monochromatic with cool bias and a warm accent", "temperature": "cool", "saturation": "desaturated", "contrast": "medium contrast" }, "lighting": { "type": "mixed", "source_count": "multiple sources", "direction": "diffused from above", "directionality": "diffused", "quality": "soft light", "intensity": "moody", "contrast_ratio": "medium contrast", "mood": "dramatic", "shadows": { "type": "soft gradual edges", "density": "gray", "placement": "under vehicle, beneath subject's coat, under jawline", "length": "short" }, "highlights": { "treatment": "subtle", "placement": "reflections on sunglasses, leather shoes, car bumper, and subtle sheen on cheekbone" }, "ambient_fill": "present", "light_temperature": "cool (blue) with a localized warm (golden) practical light" }, "technical_specs": { "medium": "digital photography", "style": "realistic", "texture": "smooth", "sharpness": "slightly soft", "grain": "film grain", "depth_of_field": "medium", "perspective": "straight on" }, "artistic_elements": { "genre": "street", "influences": [ "Paparazzi street style photography", "High fashion editorial candid shots" ], "mood": "sophisticated", "atmosphere": "Cinematic, moody urban transit with a cool, detached, fashionable energy", "visual_style": "clean" }, "typography": { "present": false, "fonts": [], "placement": "none", "integration": "none" }, "subject_analysis": { "primary_subject": "A stylish adult male walking briskly in a tailored black overcoat and sunglasses", "positioning": "center", "scale": "full", "interaction": "isolated", "facial_expression": { "mouth": "parted slightly open", "smile_intensity": "no smile", "eyes": "hidden by sunglasses, appearing to look straightforward", "eyebrows": "neutral", "overall_emotion": "serious", "authenticity": "candid" }, "hands_and_gestures": { "left_hand": "hidden entirely inside the left pocket of the overcoat", "right_hand": "hidden entirely inside the right pocket of the overcoat", "finger_positions": "not visible", "finger_interlacing": "none", "hand_tension": "relaxed", "interaction": "resting in pockets", "naturalness": "organic casual gesture" }, "body_positioning": { "posture": "walking", "angle": "facing slightly left of camera in mid-stride", "weight_distribution": "shifted onto the front right leg", "shoulders": "level" }, "hair_analysis": { "length": "medium-short, approximately 3-4 inches on top, reaching just past the nape of the neck in the back", "style_name": "slicked back undercut or taper", "texture": "straight but highly styled, wet-look texture with visible comb tracks", "imperfections": "some slight flyaway strands visible at the very back edge near the collar, natural separation of hair clumps from product wear", "part_location": "no distinct part, pushed straight back", "styling_state": "heavy product use, likely a high-hold pomade or gel giving a glossy, structured hold", "volume": "low volume, sleek and close to the scalp" } }, "background": { "setting_type": "outdoor urban street", "spatial_depth": "medium", "elements_detailed": [ { "item": "Black Jeep Wrangler (grill, bumper, front tire)", "position": "left and directly behind subject", "distance": "midground", "size": "medium", "condition": "worn", "specific_features": "matte black plastic bumper, dirty tire treads, distinct round headlight casing outline" }, { "item": "Potted plant (thin-leafed shrub/small tree, possibly oleander or olive)", "position": "left mid-upper quadrant", "distance": "background", "size": "small", "condition": "thriving", "specific_features": "housed in a dark planter, thin sprawling branches, small pale green leaves" }, { "item": "Wall sconce / lantern", "position": "top left quadrant", "distance": "background", "size": "small", "condition": "pristine", "specific_features": "black metal housing, classic trapezoidal shape, emitting warm yellow light" } ], "wall_surface": { "material": "painted wood/masonry facade", "surface_treatment": "smooth paint with architectural paneling and molding", "texture": "perfectly smooth", "finish": "matte", "color": "dark navy blue", "color_variation": "uniform", "features": "clean, horizontal routed lines, framed window with closed horizontal blinds", "wear_indicators": "pristine" }, "floor_surface": { "material": "asphalt street and concrete curb", "color": "medium slate grey", "pattern": "solid with slight urban texture/grit" }, "objects_catalog": "Subject's black overcoat, subject's grey trousers, subject's black leather pointed shoes, subject's black wayfarer sunglasses, black Jeep Wrangler (partial), potted plant, wall sconce, architectural building facade.", "background_treatment": "slightly blurred to separate subject" }, "generation_parameters": { "prompts": [ "A candid paparazzi-style street photography shot of a handsome man walking briskly down an overcast city street. He has slicked-back dark hair, wearing classic black sunglasses, a long flowing black tailored overcoat, and relaxed grey trousers. Both hands are tucked casually into his coat pockets. In the background, there is a black Jeep parked, and a dark navy blue building facade with classic molding and an illuminated warm wall sconce. Shot on 85mm lens, slight film grain, desaturated cool tones, cinematic moody lighting, fashion editorial vibe.", "A low-angle fashion street style tracking shot of a male model in mid-stride. He wears a heavy black wool trench coat, wide-leg grey dress pants, and dark wayfarer sunglasses. Wet-look slicked back hair. He walks past a dark SUV on a grey asphalt street. The background features a deep blue storefront with a glowing vintage street lamp. Soft overcast lighting, muted color palette, sharp focus on the subject with a slightly soft background." ], "keywords": [ "street style", "menswear", "candid photography", "black overcoat", "urban fashion" ], "technical_settings": "Camera: DSLR or Mirrorless, Lens: 85mm or 105mm, Aperture: f/2.8 or f/4 to achieve medium depth of field, Shutter Speed: 1/500s or faster to freeze walking motion, ISO: 400-800 for overcast lighting and slight grain.", "post_processing": "Desaturated color grade pulling down greens and warm tones (except for the localized lamp light), crushed blacks for a matte shadow look, added subtle film grain, slight vignette, cool temperature overall white balance shift." } }

Banana Pro
[image1]
人物与角色1:1

[image1] Create an Ultra-realistic cinematic dark posterdramatic over-the-shoulder pose of a young man(using the same face as uploaded image), camerapositioned slightly behind the shoulder, 90-45degree side profile facing right, strong rim lightoutlining the messy textured hair, nose and sharpjawline, deep black background, only the edges ofthe face glowing with white light, brown roundsunglasses with silver rimmed reflecting a soft glow,mysterious shadowed face with almost no facialdetails visible, intense cinematic lighting

Banana Pro
Use the uploaded CHARACTER REFERENCE SHEET as the exact MASTER image.
人物与角色3:4

Use the uploaded CHARACTER REFERENCE SHEET as the exact MASTER image. Use the uploaded REFERENCE PHOTO ONLY for the person’s facial identity. Replace ONLY the character’s face/identity with the person from the reference photo. Preserve the reference person’s recognizable facial features, face shape, eyes, nose, lips, skin texture, and natural identity. Adapt the face naturally to every angle and expression shown in the character sheet. IMPORTANT — DO NOT CHANGE ANYTHING ELSE. Keep the MASTER CHARACTER SHEET exactly the same: - Same Victorian-style floral dress - Same dress color, fabric, lace, sleeves, neckline, seams and ruffles - Same necklace and earrings - Same hairstyle and hair arrangement shown in the character sheet - Same body shape and proportions - Same full-body poses - Same front, 3/4, side and back views - Same facial expression layouts - Same costume-detail panels - Same color/material palette - Same background - Same lighting - Same composition and panel layout - Same typography and labels - Same overall image quality and realism Do NOT copy the reference person’s clothes, body, pose, accessories, background, or hairstyle. Only transfer the facial identity from the REFERENCE PHOTO. Change the character name from “Ella Beatty” to: SOFIA Keep all other character-sheet information unchanged unless it directly needs to be adjusted for visual consistency. Make the face replacement seamless and photorealistic, with accurate facial proportions, natural skin texture, realistic lighting, and consistent identity across every portrait, expression, pose, and turnaround view. FINAL RULE: REFERENCE PHOTO = FACE/IDENTITY ONLY. MASTER CHARACTER SHEET = EVERYTHING ELSE. Do not redesign or recreate the character sheet. Preserve the original sheet and modify only the requested face identity and the character name “Sofia”. Image: 2 Use the uploaded TARGET IMAGE as the exact master image. Use the uploaded REFERENCE IMAGE only for the person’s identity: face, facial features, hairstyle, hairline, hair texture, beard/facial hair, skin tone, and overall identity. IMPORTANT: Do NOT copy the reference person’s clothing, outfit, accessories, body pose, background, or composition. Replace the face and identity of the person in the TARGET IMAGE with the person from the REFERENCE IMAGE. Match the reference person’s face, hair, beard/facial hair, and identity naturally and accurately to the target’s head angle, lighting, perspective, and expression. Keep EVERYTHING ELSE from the TARGET IMAGE exactly unchanged: - Same clothing/outfit - Same shirt, trousers, shoes, and accessories - Same body proportions and physique - Same pose and hand position - Same background - Same environment - Same lighting and shadows - Same camera angle - Same framing and composition - Same image quality and realism The TARGET IMAGE clothing must remain exactly as shown in the target. NEVER use the reference person’s clothes. Create a photorealistic, seamless face replacement with natural skin texture, realistic facial proportions, accurate hair/beard integration, and matching lighting. FINAL RULE: REFERENCE IMAGE = identity only. TARGET IMAGE = clothing + body + pose + background + composition + everything else. Do not change anything except the person’s identity/face, hair, and facial hair according to the reference image.

GPT Image 2.5
Create a vertical two-panel editorial artwork using the uploaded photograph as the exac...
人物与角色2:3

Create a vertical two-panel editorial artwork using the uploaded photograph as the exact visual reference. TOP HALF — ORIGINAL PHOTO Keep the top half as the real, unaltered photograph. Preserve the subject's identity, face, hair, clothing, pose, proportions and gesture. For animals, preserve coat colour, markings, ears, body shape, tail and action. If there is a person-and-animal pair, preserve their exact relationship and positioning. Architecture and complex background details remain photographic only and must not be recreated in the lower half. BOTTOM HALF — NATURAL-MATERIAL RECREATION Rebuild the same subject, pose and silhouette as a sophisticated editorial illustration made entirely from real leaves, petals, twigs, grass stems, fruits, seeds, bark and dried leaves. The natural materials must form the actual structure of the subject—not simply decorate a drawn figure. Use their natural shapes intelligently: - Hair from layered or curling leaves and grass. - Clothing from overlapping leaves and petals. - Arms and legs from twigs, grass stems or narrow leaves. - Ears from naturally pointed leaves. - Tails from curved twigs or slender plant material. - Face and body from the natural contours, textures and edges of the materials. Do not draw the subject's outline first and fill it with plants. Do not cut plants into geometric shapes or create a children's leaf-craft appearance. The same person, animal or pair from the photograph must remain immediately recognisable. COMPOSITION & COLOUR Place the natural-material subject within one simple rectangular colour field on a pale paper background. The rectangle must have a perfectly straight bottom and right edge, with no notches, steps or inner panels. Allow only a few naturally directional elements—such as a hand, ear, tail, twig or grass stem—to extend beyond the field organically. Keep generous white space, an elegant centre of gravity and slightly imperfect symmetry. Match the photograph's strongest colours using the natural colours of the materials themselves. Preserve real variations in leaf, petal, fruit, bark and dried-plant colours. Never make all materials artificially uniform. Use black linework only as a minimal supporting detail for eyes, mouth, tiny motion marks or small handwritten accents. Never create a complete black outline around the subject. TEXT & TYPOGRAPHY Include exactly three pieces of text: 1. One short English title. 2. One tiny annotation naming a real plant, material or object. 3. One different tiny annotation naming another real plant, material or object. All text must use Latin letters only, be correctly spelled, and contain no copied signage, logos or text from the original photograph. No word may be repeated between the three text elements. Keep typography subtle and editorial, never like a conventional poster headline. EMPTY BOTTOM AREA The bottom sixth of the entire image must remain completely empty paper, except the title if needed. No plants, colour field, twig, annotation or drawn line may enter this area. The right third of this bottom strip must be completely empty, including the title. FINAL LOOK High-end contemporary editorial art direction: real photograph × real natural materials × delicate hand-drawn trace × clean geometric colour field × restrained typography. The result should feel poetic, intelligent, handmade and intentionally art-directed—as if an artist physically selected natural materials and reconstructed the photographed subject from their own organic forms. Do not invent subjects, duplicate limbs/objects, alter the identity, use complete outlines, create cartoon aesthetics, or turn the architecture into the natural-material collage.

GPT Image 2.5
A photorealistic image on the left displays a young woman standing in a narrow alleyway...
人物与角色3:2

A photorealistic image on the left displays a young woman standing in a narrow alleyway lined with traditional buildings and potted plants, wearing a white top, a muted green cardigan, and dark jeans while carrying a brown leather tote bag and thoughtfully touching her chin. Beside it on the right is a stylized watercolor illustration capturing the exact same scene, featuring loose ink outlines and soft color washes on aged, textured paper. The illustration includes artistic script on the left side reading \"Another Path Dreamer Scenes\" and text at the bottom reading \"A QUIET MOMENT,\" effectively transforming the real-world photograph into a whimsical, storybook-style artistic rendering.

GPT Image 2.5
从一个以人为中心的生活瞬间创作一张9:16的编辑排版海报。
人物与角色2:3

从一个以人为中心的生活瞬间创作一张9:16的编辑排版海报。\n使用超大排版作为主要构图结构。让人、动作、城市空间或日常场景出现在字母内部、穿过字母或突破字母边缘。图像应与文字融为一体,而不仅仅是放置在文字后面。主题:[自定义标题]\n主体:[主体性别或来源]\n人类状态:[人类状态]\n调色板:[调色板]\n系列:人类碎片\n年份:2026

GPT Image 2
Caption:One mind. Infinite versions. Which one are you choosing?
人物与角色2:3

Caption:One mind. Infinite versions. Which one are you choosing?\nPrompt:\nCreate an ultra-realistic, high-fashion editorial photograph inside a luxurious minimalist studio with smooth light-gray walls, polished reflective concrete floor, and large floor-to-ceiling windows casting soft natural sunlight. At the center, a confident young man with neatly styled black hair, a trimmed beard, and black rectangular sunglasses sits casually in a modern black designer chair with wooden legs, wearing an oversized all-white streetwear outfit, ripped white jeans, white sneakers, a black smartwatch, and a relaxed confident expression. Around him are multiple miniature versions of himself (8–10 tiny clones), each in a different pose—walking, sitting, leaning, standing with hands in pockets, adjusting sunglasses, relaxing on the chair arm, sitting on the floor, and one being delicately held between giant fingers entering from the top-left of the frame. The tiny clones interact naturally with the environment, creating a playful forced-perspective illusion. Clean luxury interior, dramatic scale contrast, soft shadows, realistic reflections, premium fashion photography, ultra-detailed facial features, cinematic composition, depth of field, volumetric daylight, hyper-realistic textures, editorial magazine quality, photorealistic, 8K, HDR, masterpiece.

GPT Image 2
Lumina Mascot LogoLumina · AI 提示词与光影画廊·纯粹呈现 · 非营利性灵感交流与提示词研究
免责声明