Infographic / Edu Visual - Dark Research Figure Gallery UI

Creates a dark web gallery interface displaying curated research paper figure style cards for an academic visual preset library.

提示词

3,561 字符打开 Astorie 创作台 ↗
Goal: Create a dark, high-end gallery index page for a curated preset library of research paper figure styles, titled {argument name="headline text" default="研究论文图示"}. The interface should look like a polished web app screenshot showcasing academic diagram thumbnails.

Canvas: Wide landscape browser-like composition, approximately 16:9, dark charcoal background, subtle grid borders, crisp UI typography, high contrast white and cyan text.

Header: Center the title at the top with a small stacked multicolor document icon to its left. Add a thin horizontal divider line below the header. In the upper-right area, place a compact rounded button labeled “↑ 图库索引”.

Section title: On the left below the divider, show the section heading {argument name="section title" default="研究论文图示网格"} in bold light text.

Layout: Display a 4-column gallery grid with thin gray borders and dark card gutters. Show exactly 12 visible gallery cards: 8 full cards in the first two rows and 4 partially visible cards in the third row at the bottom edge. Each card contains a white research-paper-style thumbnail image, a bold lettered title underneath, and three small dark rounded tag pills below the title reading exactly “landscape”, “high”, and “Curated”.

Visible card count and labels: Card A: “患者队列与多模态生物标志物流称流程”, thumbnail showing a patient cohort workflow with boxes, charts, and a survival curve. Card B: “单细胞免疫图谱”, thumbnail showing a single-cell immune atlas with UMAP clusters, dot plot, stacked bar chart, and trajectory plot. Card C: “多模态医疗 AI 方法图”, thumbnail showing a multimodal foundation model clinical decision support diagram with medical image, pathology, text, and model blocks. Card D: “治疗响应统计图”, thumbnail showing therapeutic response statistics with bar charts, forest plots, scatter plot, and a circular workflow. Card E: “Transformer 编码器-解码器架构”, thumbnail showing a transformer encoder-decoder architecture diagram with stacked module blocks. Card F: “多智能体 LLM 系统架构”, thumbnail showing an LLM multi-agent system architecture with central model block, surrounding tool icons, memory, reflection, and evaluation modules. Card G: “去噪扩散正/逆向链”, thumbnail showing a denoising diffusion forward and reverse process chain with noisy image panels and arrows. Card H: “经验缩放规律图”, thumbnail showing empirical scaling laws with multiple descending colored curves and a legend. Bottom partially visible card 1: title begins “Benchmark comparison across 10 frontier LLMs”, thumbnail with bar chart comparisons. Bottom partially visible card 2: title begins “Ablation of core reasoning components across 5 benchmarks”, thumbnail with grouped bars. Bottom partially visible card 3: title begins “LLM pretraining data mixture and downstream splits”, thumbnail with stacked area/data mixture blocks. Bottom partially visible card 4: title begins “Representative multi-head attention patterns in a 16-layer Transformer”, thumbnail with heatmaps.

Visual style: Use a modern SaaS dashboard aesthetic, dark mode, neat academic curation feel, subtle cyan accents, small but legible labels, and realistic research figure thumbnails that resemble high-quality journal diagrams. Keep thumbnails varied but consistently white-background scientific figures.

Constraints: Preserve the Chinese interface labels and card titles as written. Do not add people, photos, watermarks, browser address bars, or decorative clutter. The composition should feel like a screenshot of a curated research-figure template library named {argument name="project name" default="GPT-IMAGE-2-SKILL"}.
来自原帖查看来源 ↗

更多灵感

Social Media Post - Haaland Anime Confrontation

Social Media Post - Haaland Anime Confrontation

Popcraft AIGPT Image 2
A dramatic anime-style illustration of Erling Haaland in a Manchester City jersey, screaming furiously with his hands gripping his hair. He is confronting a bald referee who has his back turned to the viewer, holding up a hand to calm him down. Another player is partially visible on the right side. The scene takes place during a rainy match, with water droplets falling and stadium lights creating a blurred background.
Social Media Post - GPT Image 2 Notebook Pop-Out Illusion

Social Media Post - GPT Image 2 Notebook Pop-Out Illusion

AizaGPT Image 2
Create an ultra-detailed photorealistic scene of a young woman seemingly emerging through a torn sheet of lined spiral notebook, a convincing 3D pop-out optical-illusion; slight top-down angled perspective showing the spiral binding on the left and wooden desk surface around the notebook, soft natural warm daylight illuminating her face, shallow depth of field, 35mm perspective, f/1.8, crisp focus on her eyes. She has long blonde hair in soft waves, sunglasses perched on her head, natural rosy makeup, gentle smile, wearing a cream cable-knit sweater and a beige leather-strap watch; realistic skin texture and pores, subtle freckles, lifelike catchlights. Torn paper edges are ragged and curled with scattered paper scraps and fibers, strong tactile paper shadows and torn-cardboard thickness visible. A right hand holding a sharpened pencil sketches or traces the torn edge from the top-right, another hand reaches through the hole toward the viewer. Nearby on the wooden desk: black scissors and a yellow utility knife with metal blade, soft specular highlights. Hyper-real textures on paper, fabric, wood and metal, accurate subsurface scattering on skin, warm cinematic color grading, ultra-high resolution, photoreal render, finely detailed lighting, realistic shadows, subtle film grain.
Social Media Post - GPT Image 2 Pastel Crayon Illustration

Social Media Post - GPT Image 2 Pastel Crayon Illustration

TaarukGPT Image 2
Transform the uploaded reference photo into a charming, hand-drawn pastel crayon illustration while preserving the original subject’s exact identity, facial features, expression, pose, proportions, clothing or accessories, and overall composition. Keep the pet exactly recognizable, including its fur color, eyes, body shape, and distinctive details. Recreate the image in a cute, nostalgic children’s storybook art style using soft wax crayons, colored pencils, and textured pastel strokes. Use a warm cream-colored textured paper background with visible paper grain, rough hand-drawn outlines, imperfect sketchy edges, and soft layered crayon shading. Add small playful hand-drawn doodles around the subject, such as hearts, flowers, stars, sunshine, grass, tiny decorative illustrations, and cute themed elements that complement the original scene. Include a few short handwritten phrases in a casual, childlike style to create a cozy scrapbook aesthetic. Use a soft, warm pastel color palette with natural earthy tones. Keep the subject large and central, with the same camera angle and framing as the original photograph. The final result should look like a beautiful handmade crayon drawing on vintage sketchbook paper, adorable, warm, playful, and full of personality. Do not change the subject or redesign the scene; only transform the photographic image into this illustrated art style.
Profile / Avatar - Bedroom Selfie Portrait Prompt

Profile / Avatar - Bedroom Selfie Portrait Prompt

沧瑶GPT Image 2
【Overall Style】

1:1 square composition, single independent frame, ultra-realistic Korean-style adult female indoor portrait photography, gentle sweet girlish feeling, cream pink tones, realistic close-range mobile phone selfie, soft natural light, slight soft focus, retaining real photography texture.

【Facial Features】

A beautiful adult East Asian female aged 23 to 25, delicate small oval face, soft smooth jawline, small rounded chin, clear fine cold white skin.

Bright clear deep brown round almond eyes, natural double eyelids, soft full lying silkworms, slender curled eyelashes, fine light pink-brown eye makeup.

Slender straight nose bridge, small nose tip, cheeks with distinct but naturally blended peach pink blush.

Soft full cherry blossom rose pink lips, delicate moist glass lip luster.

Lips naturally slightly parted, eyes gently staring directly at lens, expression quiet, sweet, slightly lazy, retaining natural skin micro-texture.

【Hairstyle - Core】

Extremely dense and fluffy deep black tea color waist-length long hair, with natural deep chocolate brown luster.

Top of head naturally fluffy, adopting center-parted hairstyle, forehead has light messy air broken bangs.

Both sides long hair naturally draping over shoulders and chest, with soft loose large arc waves.

Few slender hair strands naturally sweeping across forehead and cheeks, rich hair layers, retaining real independent hair strands and natural messiness.

【Clothing & Jewelry】

Wearing shallow cherry blossom pink thin strap camisole top, soft cotton elastic fabric, simple circular arc neckline.

Outer layer cream white loose knit cardigan, fine soft knit texture, front placket naturally open, long sleeves slightly covering wrists.

Wearing small white pearl earrings and extremely thin silver clavicle chain.

【Character Pose】

Character sitting on bedroom bed edge, upper body naturally slightly leaning forward, body front facing lens.

Head gently tilted, chin slightly tucked, eyes directly staring at lens.

Hand on left side of frame raised to hair side, fingers naturally curved, gently organizing long hair near ear.

Other hand gently supporting cheek, palm naturally fitting jaw edge, fingers softly stretched.

Hands posture relaxed and natural, presenting gentle intimate selfie atmosphere.

【Background Environment】

Warm simple pink-tone bedroom, behind character is a large bed covered with shallow cherry blossom pink sheets, pink pillows and cream white bedding.

Background is light pink beige walls and simple headboard, overall environment soft, clean, cozy.

【Lens & Lighting】

1:1 square, equivalent 35 to 50mm lens, close-range chest-up half-body close-up, eye-height level view.

Character face located center of frame, top of head close to top edge of frame, two hands naturally surrounding cheek composition.

Soft cold white frontal fill light evenly illuminating face, lips and hands, skin tone clear and natural.

Background retains faint warm pink ambient light, forming soft warm-cold contrast.

Low saturation cream pink tones, delicate real skin quality, soft highlights, slight camera grain, no excessive smoothing, no excessive sharpening.

【Negative Prompts】

Anime, illustration, 3D rendering, CG fake face, plastic skin, excessive smoothing, abnormally large eyes, overly pointed chin, facial feature misalignment, wig, matted hair, stiff curls, extra fingers, deformed palms, twisted wrists, abnormal body proportions, clothing clipping, severe overexposure, strong HDR, over-sharpening, obvious AI feel, text, watermark, LOGO, collage.
Product Marketing - Classic Cafe Date Outfit Prompt

Product Marketing - Classic Cafe Date Outfit Prompt

のぞむ*AIイラストGPT Image 2
OUTFIT:
A {argument name="dress color" default="navy blue"} satin midi dress with elegant long sleeves and a delicate white lace collar, creating a refined, classic silhouette.

Accented with a thin pearl belt at the waist and layered with a small white bolero jacket for a graceful feminine touch.

Styled with black Mary Jane heels decorated with small ribbon details and a compact pearl clutch bag.

Elegant, romantic, classic café-date fashion aesthetic.
Navy blue, crisp white, black, and pearl ivory color palette.
Smooth satin, delicate lace, luminous pearls, and polished fabric textures.
Sophisticated, feminine, timeless seasonal styling.
Social Media Post - Watercolor Illustration Comparison Poster

Social Media Post - Watercolor Illustration Comparison Poster

SairahGPT Image 2
Transform the uploaded photograph into a beautiful, hand-drawn watercolor illustration comparison poster.

LAYOUT & COMPOSITION:
Create a vertical 3:4 poster divided into two equal horizontal sections.

TOP HALF — ORIGINAL PHOTO:
Display the original uploaded photograph exactly as it is. Preserve the original subject, pose, clothing, background, lighting, colors, and composition without alterations.

BOTTOM HALF — WATERCOLOR ILLUSTRATION:
Recreate the exact same photograph as a delicate, hand-painted watercolor and colored-pencil illustration. Preserve the subject's identity, facial features, hairstyle, expression, clothing, pose, proportions, and surroundings.

Use soft watercolor washes, delicate ink outlines, visible pencil strokes, subtle paper texture, warm pastel tones, gentle shading, and beautifully imperfect hand-drawn details. Give the illustration a charming Korean/Japanese lifestyle sketchbook aesthetic.

DECORATIVE DETAILS:
Add handwritten playful typography that complements the subject and scene. Include small hand-drawn doodles such as hearts, stars, arrows, crowns, underlines, or sparkles. Generate a short, creative phrase that matches the subject's personality or mood.

BACKGROUND:
Maintain the original environment but reinterpret it with loose watercolor brushstrokes, simplified scenery, and soft artistic details.

OVERALL STYLE:
Dreamy, cozy, youthful, elegant, artistic, and slightly whimsical. Use warm off-white textured watercolor paper as the illustration background.

IMPORTANT:
- Keep the original photograph completely unchanged in the top section.
- The bottom illustration must clearly depict the exact same moment.
- Do not invent additional people or objects.
- Maintain strong visual similarity to the original photograph.
- Make the typography look naturally handwritten, not digitally typeset.
- Ensure a clean 50/50 split with balanced composition.

High-quality editorial art poster, 3:4 vertical aspect ratio.

寻找下一个方向

ASTORIE CANVAS

让灵感,继续生长。

在同一个创作空间,连接提示词、参考素材和生成结果。

打开创作画布 ↗