Skip to content

karlorz/react-image-annotate

 
 

Repository files navigation

React Image Annotate

npm version Ask DeepWiki

The best image/video annotation tool ever. Check out the demo here. Or the code sandbox here.

Features

  • Simple input/output format
  • Bounding Box, Point and Polygon Annotation
  • Zooming, Scaling, Panning
  • Multiple Images
  • Cursor Crosshair
  • Headless Architecture (v3.x) - Build custom UIs with useAnnotator hook
  • Internationalization (i18n) - Built-in support for EN/ZH/VI languages
  • Theme Customization - PingFang font stack with Material-UI v6

Screenshot of Annotator

Usage

Install with npm (recommended):

npm install @karlorz/react-image-annotate

Or with bun (alternative):

bun add @karlorz/react-image-annotate
import React from "react";
import ReactImageAnnotate from "@karlorz/react-image-annotate";

const App = () => (
  <ReactImageAnnotate
    labelImages
    regionClsList={["Alpha", "Beta", "Charlie", "Delta"]}
    regionTagList={["tag1", "tag2", "tag3"]}
    images={[
      {
        src: "https://placekitten.com/408/287",
        name: "Image 1",
        regions: []
      }
    ]}
  />
);

export default App;

Headless Architecture (v3.x)

For custom UIs, use the headless useAnnotator hook:

import { useAnnotator } from "@karlorz/react-image-annotate/headless";

function MyCustomAnnotator() {
  const { state, actions, regions, currentImage } = useAnnotator({
    images: [{ src: '/image.jpg', name: 'Test' }],
    onExit: (output) => console.log(output)
  });

  return (
    <div>
      <button onClick={() => actions.selectTool('create-box')}>
        Box Tool
      </button>
      <canvas {...actions.getCanvasProps()} />
      {regions.map(r => (
        <div key={r.id}>
          {r.cls}
          <button onClick={() => actions.deleteRegion(r)}>Delete</button>
        </div>
      ))}
    </div>
  );
}

See examples/headless/ for complete examples with Tailwind CSS, Semi Design, and vanilla React.

Internationalization (i18n)

Wrap your app with I18nProvider for multi-language support:

import { I18nProvider } from "@karlorz/react-image-annotate";

<I18nProvider language="zh"> {/* en, zh, or vi */}
  <ReactImageAnnotate {...props} />
</I18nProvider>

Or provide custom translations:

<I18nProvider
  language="fr"
  translations={{
    fr: { 'header.save': 'Enregistrer', ... }
  }}
>
  <ReactImageAnnotate {...props} />
</I18nProvider>

Theme Support (Light/Dark Mode)

The component supports both light and dark themes:

// Simple string mode (recommended)
<ReactImageAnnotate theme="dark" {...props} />
<ReactImageAnnotate theme="light" {...props} />

// Custom Material-UI theme
import { createTheme } from '@mui/material/styles'

const customTheme = createTheme({
  palette: {
    mode: 'dark',
    primary: { main: '#3f51b5' },
    background: {
      default: '#1a1a1a',
      paper: '#2a2a2a',
    },
  },
})

<ReactImageAnnotate theme={customTheme} {...props} />

Theme Integration: The component respects the MUI theme palette and will automatically style all internal components (workspace, sidebars, dialogs) according to the theme mode. Perfect for applications with dynamic theme switching!

To get the proper fonts, make sure to import the Inter UI or Roboto font, the following line added to a css file should suffice.

@import url("https://rsms.me/inter/inter.css");

Props

All of the following properties can be defined on the Annotator...

Prop Type (* = required) Description Default
taskDescription *string Markdown description for what to do in the image.
allowedArea { x: number, y: number, w: number, h: number } Area that is available for annotation. Entire image.
regionTagList Array<string> Allowed "tags" (mutually inclusive classifications) for regions.
regionClsList Array<string> Allowed "classes" (mutually exclusive classifications) for regions.
imageTagList Array<string> Allowed tags for entire image.
imageClsList Array<string> Allowed classes for entire image.
enabledTools Array<string> Tools allowed to be used. e.g. "select", "create-point", "create-box", "create-polygon" Everything.
showTags boolean Show tags and allow tags on regions. true
selectedImage string URL of initially selected image.
images Array<Image> Array of images to load into annotator
showPointDistances boolean Show distances between points. false
pointDistancePrecision number Precision on displayed points (e.g. 3 => 0.123)
theme "light" | "dark" | Theme Theme mode (string) or full MUI theme object for styling the component. "light"
onExit MainLayoutState => any Called when "Save" is called.
RegionEditLabel Node React Node overriding the form to update the region (see RegionLabel)
allowComments boolean Show a textarea to add comments on each annotation. false
hidePrev boolean Hide Previous Image button from the header bar. false
hideNext boolean Hide Next Image button from the header bar. false
hideClone boolean Hide Clone button from the header bar. false
hideSettings boolean Hide Settings button from the header bar. false
hideFullScreen boolean Hide FullScreen/Window button from the header bar. false
hideSave boolean Hide Save button from the header bar. false

Developers

Development

This project uses Storybook with Vite. To begin developing, run the following commands in the cloned repo:

  1. npm install (or bun install)
  2. npm run storybook (or bun run storybook)

A browser tab will automatically open with the project components at http://localhost:9090.

Package Manager

  • Primary: npm (recommended)
  • Alternative: bun (faster builds)

See more details in the contributing guidelines.

Icons

Consult these icon repositories:

About

Create image annotations. Classify, tag images with polygons, bounding boxes or points.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • JavaScript 93.4%
  • CSS 4.8%
  • TypeScript 1.6%
  • HTML 0.2%