Background Removal in the Browser

We are always looking for great people at IMG.LY. If you are working with our background removal library you might be a perfect fit! Apply now at IMG.LY Careers

<p align="center"> <img src="https://img.ly/showcases/cesdk/web/s/case-thumbnail/background-removal/background-removal-0.png?utm_source=github&utm_medium=project&utm_campaign=background-removal-js" alt="background removal js showcase" /> </p>

Remove backgrounds from images directly in the browser environment with ease and no additional costs or privacy concerns. Explore an interactive demo.

News

For more detail information please refer to the CHANGELOG.

Overview

@imgly/background-removal is a powerful npm package that allows developers to seamlessly remove the background from images directly in the browser. With its unique features and capabilities, this package offers an innovative and cost-effective solution for background removal tasks without compromising data privacy.

The key features of @imgly/background-removal are:

• In-Browser Background Removal: Our one-of-a-kind solution performs the entire background removal process directly in the user's browser, eliminating the need for additional server costs. By leveraging the computing power of the local device, users can enjoy a fast and efficient background removal process.

• Data Protection: As @imgly/background-removal runs entirely in the browser, users can have peace of mind knowing that their images and sensitive information remain secure within their own devices. With no data transfers to external servers, data privacy concerns are effectively mitigated.

• Seamless Integration with IMG.LY's CE.SDK: @imgly/background-removal provides seamless integration with IMG.LY's CE.SDK, allowing developers to easily incorporate powerful in-browser image matting and background removal capabilities into their projects.

The Neural Network (ONNX model) and WASM files used by @imgly/background-removal are hosted by IMG.LY by default. See the section Custom Asset Serving if you want to host them on your own servers.

Installation

You can install @imgly/background-removal via npm or yarn. Use the following commands to install the package:

NPM

npm install @imgly/background-removal

Usage

import imglyRemoveBackground from "@imgly/background-removal"

let image_src: ImageData | ArrayBuffer | Uint8Array | Blob | URL | string = ...;

imglyRemoveBackground(image_src).then((blob: Blob) => {
  // The result is a blob encoded as PNG. It can be converted to an URL to be used as HTMLImage.src
  const url = URL.createObjectURL(blob);
})

Note: On the first run the wasm and onnx model files are fetched. This might, depending on the bandwidth, take time. Therefore, the first run takes proportionally longer than each consecutive run. Also, all files are cached by the browser and an additional model cache.

Advanced Configuration

The library does not need any configuration to get started. However, there are optional parameters that influence the behaviour and give more control over the library.

type Config = {
  publicPath: string; // The public path used for model and wasm files. Default: 'https://staticimgly.com/${PACKAGE_NAME}-data/${PACKAGE_VERSION}/dist/'
  debug: bool; // enable or disable useful console.log outputs
  device: 'cpu' | 'gpu'; // choose the execution device. gpu will use webgpu if available
  proxyToWorker: bool; // Whether to proxy the calculations to a web worker. (Default true)
  model: 'isnet' | 'isnet_fp16' | 'isnet_quint8'; // The model to use. (Default "isnet_fp16")
  output: {
    format: 'image/png' | 'image/jpeg' | 'image/webp'; // The output format. (Default "image/png")
    quality: number; // The quality. (Default: 0.8)
    type: 'foreground' | 'background' | 'mask'; // The output type. (Default "foreground")
  };
};

Download Size vs Quality

The onnx model is shipped in various sizes and needs.

• small (~40 MB) is the smallest model and is in most cases working fine but sometimes shows some artifacts. It's a quantized model.
• medium (~80MB) is the default model.

Preloading Assets

Per default, assets are downloaded on demand. You might enforce downloading all assets at once at any time by preloading them:

import removeBackground, { preload, Config } from '@imgly/background-removal';

const config: Configuration = ...;
preload(config).then(() => {
  console.log("Asset preloading succeeded")
})

Download Progress Monitoring

On the first run, the necessary data will be fetched and stored in the browser cache. Since the download might take some time, you have the option to tap into the download progress.

let config: Config = {
  progress: (key, current, total) => {
    console.log(`Downloading ${key}: ${current} of ${total}`);
  }
};

Performance

The performance is largely dependent on the feature set available. Most prominently, ensure that SharedArrayBuffer is available MDN. Due to the security requirements of SharedArrayBuffer two headers need to be set to cross-origin isolate your site:

'Cross-Origin-Opener-Policy': 'same-origin',
'Cross-Origin-Embedder-Policy': 'require-corp'

Custom Asset Serving

The wasm and onnx neural networks are hosted by IMG.LY by default. For production use, we advise you to host them yourself:

• Download the following package with the package version that matches your @imgly/background-removal version from the IMG.LY CDN and decompress it. Note that you need to replace YOUR_PACKAGE_VERSION with the actual version of the package you are using. The URL is https://staticimgly.com/@imgly/background-removal-data/YOUR_PACKAGE_VERSION/package.tgz.
• Move the content of the package/dist folder to be served by your web server. This often is the /public folder.

import imglyRemoveBackground, {Config} from "@imgly/background-removal"

const public_path = "https://example.com/assets/" ; // the path assets are served from

let config: Config =  {
  publicPath: public_path, // path to the wasm files
};

let image_src: ImageData | ArrayBuffer | Uint8Array | Blob | URL | string = ...;

imglyRemoveBackground(image_src, config).then((blob: Blob) => {
  // result is a blob encoded as PNG.
  // It can be converted to an URL to be used as HTMLImage.src
  const url = URL.createObjectURL(blob);
})

Debug Outputs

Enable debug outputs and logging to the console

let config: Config = {
  debug: true
};

Cross-Origin Resource Sharing (CORS)

If you are running into CORS issues you might want to pass additional parameters to the fetch function via

let config: Config = {
  fetchArgs: {
    mode: 'no-cors'
  }
};

Fetch Args

fetchArgs are passed as second parameters to the fetch function as described in MDN.

Who is it for?

@imgly/background-removal is ideal for developers and projects that require efficient and cost-effective background removal directly in the browser. It caters to a wide range of use cases, including but not limited to:

• E-commerce applications that need to remove backgrounds from product images in real time.

• Image editing applications that require background removal capabilities for enhancing user experience.

• Web-based graphic design tools that aim to simplify the creative process with in-browser background removal.

Whether you are a professional developer or a hobbyist, @imgly/background-removal empowers you to deliver impressive applications and services with ease.

License

The software is free for use under the AGPL License. Please contact support@img.ly for questions about other licensing options.

Authors & Contributors

This library is made by IMG.LY shipping the world's premier SDKs for building creative applications. Start your trial of the CreativeEditor SDK, PhotoEditor SDK & VideoEditor SDK.