What Are the 12 Visual Categories in the GPT-Image2 Case Gallery?

The GPT-Image2 case gallery organizes 544 generated image examples into 12 distinct visual categories—including UI design, infographics, photography, and illustration—each accessible via anchored sections in the freestylefly/awesome-gpt-image-2 repository.

The freestylefly/awesome-gpt-image-2 project maintains a comprehensive reference gallery demonstrating GPT-Image2 capabilities across diverse design disciplines. These 12 visual categories serve as the primary taxonomy for navigating the collection, with each category defined by specific anchor tags in the documentation and corresponding metadata in the project's data layer.

The 12 Visual Categories in GPT-Image2

The gallery classifies its 544 cases into the following visual disciplines, ranging from interface design to historical art styles:

UI and Interface (UI 与界面)

73 cases covering application mockups, dashboard layouts, and live-stream interface designs. This is the third-largest category in the collection. Navigate to this section using the anchor #cat-ui in docs/gallery.md.

Charts and Information Visualization (图表与信息可视化)

53 cases of data-driven visuals including infographics, knowledge graphs, and statistical charts. Marked by the anchor #cat-infographic.

Posters and Typography (海报与排版)

90 cases making this the largest category, encompassing typographic layouts, campaign artwork, and print-ready poster designs. Access via #cat-poster.

Products and E-commerce (商品与电商)

42 cases featuring product showcase images, e-commerce listing photos, and digital advertising creatives. Anchor: #cat-product.

Brand and Logo Design (品牌与标志)

27 cases dedicated to brand identity systems, corporate logos, and mascot character designs. Anchor: #cat-brand.

Architecture and Space (建筑与空间)

12 cases of architectural renders, interior design visualizations, and space-planning diagrams. Anchor: #cat-architecture.

Photography and Realism (摄影与写实)

78 cases of photorealistic imagery including portrait shoots, lifestyle photography, and realistic still-life compositions. This is the second-largest category. Anchor: #cat-photo.

Illustration and Art (插画与艺术)

59 cases spanning artistic sketches, stylized digital paintings, and conceptual artwork. Anchor: #cat-illustration.

Characters and Roles (人物与角色)

31 cases focusing on character design sheets, avatar collections, and role-play illustration sets. Anchor: #cat-character.

Scenes and Narrative (场景与叙事)

21 cases of narrative-driven visuals including storyboard sequences and cinematic concept art. Anchor: #cat-scene.

History and Ancient Style (历史与古风题材)

16 cases of "gu-feng" (ancient-style) visualizations depicting historical settings and traditional aesthetic themes. Anchor: #cat-history.

Documents and Publications (文档与出版物)

11 cases representing the smallest category, containing magazine layouts, book page designs, and document formatting templates. Anchor: #cat-document.

Note: The repository also includes a miscellaneous Other Application Scenarios (其他应用场景) category with 28 cases (anchor #cat-other), serving as a catch-all for cross-disciplinary applications that do not fit the primary 12 visual taxonomies.

How Categories Are Structured in the Source

The category taxonomy is formally defined in docs/gallery.md, where each visual discipline uses an HTML anchor tag followed by an H3 header. For example, the Photography category begins with:

<a name="cat-photo"></a>

### 摄影与写实· 78 cases

The structured metadata for all 544 examples resides in data/cases.json, which includes category assignments for programmatic access. The React frontend entry point at src/main.jsx consumes this JSON to render the categorized gallery interface, while src/supabaseClient.js manages user-specific interactions such as favorites that may be filtered by these visual categories.

Programmatically Accessing Category Data

You can extract the complete category list directly from the markdown documentation using Node.js:

import fs from 'fs';
import path from 'path';

const galleryPath = path.resolve('docs/gallery.md');
const content = fs.readFileSync(galleryPath, 'utf-8');

// Capture all category anchors and titles
const catRegex = /<a name="cat-(.+?)"><\/a>\s*\n\s*###\s*([^·]+)·\s*\d+ cases/g;
let match;
const categories = [];

while ((match = catRegex.exec(content)) !== null) {
  const slug = match[1];
  const title = match[2].trim();
  categories.push({ slug, title });
}

console.log(categories);

Executing this script returns an array of objects mapping slugs to Chinese category titles (e.g., {slug: 'ui', title: 'UI 与界面'}), enabling automated categorization workflows or static site generation based on the gallery's visual taxonomy.

Summary

• The 12 visual categories in the GPT-Image2 case gallery span interface design, data visualization, commercial photography, illustration, architecture, and specialized domains like ancient-style art.
• Categories are anchored in docs/gallery.md using HTML anchor tags (e.g., #cat-infographic) for direct URL navigation to specific sections.
• The repository contains 544 total examples, with Posters and Typography (90 cases) representing the largest classification and Documents (11 cases) the smallest.
• Case metadata is stored in data/cases.json and rendered via the React frontend in src/main.jsx, with src/supabaseClient.js handling user-data persistence.

Frequently Asked Questions

What is the largest visual category in the GPT-Image2 gallery?

Posters and Typography (海报与排版) contains 90 cases, making it the most populous category. It is followed by Photography and Realism with 78 cases, and UI and Interface with 73 cases.

How can I link directly to a specific category in the documentation?

Each category uses a named HTML anchor in docs/gallery.md. Append the anchor hash to the file URL—for example, #cat-illustration for the Illustration category or #cat-architecture for architectural renders.

Are the category names available in English?

The canonical category identifiers use Chinese nomenclature (e.g., 图表与信息可视化, 摄影与写实) as implemented in the source markdown and JSON data structure. The English descriptions provided here translate these for reference, but the repository's internal logic relies on the Chinese titles and slug identifiers (e.g., cat-photo, cat-ui).

How does the frontend organize cases by category?

The application entry point at src/main.jsx reads data/cases.json, which contains category assignments for each of the 544 cases. The Supabase client in src/supabaseClient.js may additionally filter or enhance these categories with user-generated data such as favorites or custom collections.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}