README.md

webpack-face-crop-loader

Getting Started

To begin, you'll need to install webpack-face-crop-loader:

$ npm install webpack-face-crop-loader --save-dev

Then add the loader to your webpack config. For example:

file.js

import profile from "./pictures_dir/profile.jpg";

webpack.config.js

module.exports = {
  module: {
    rules: [
      {
        test: /\.(png|jpe?g|gif)$/i,
        include: [path.resolve(__dirname, "pictures_dir")],
        use: [
          {
            loader: "webpack-face-crop-loader",
            options: { ...options },
          },
        ],
      },
    ],
  },
};

And run webpack via your preferred method.

Options

height

Type: number Default: 400

The height of the output image.

width

Type: number Default: 400

The width of the output image.

minScale

Type: number Default: 1

The minimal scale of the crop rect, set to 1 to prevent smaller than necessary crops. Lower than 1 will zoom in if needed (more risk of chopping things off).

inputSize

Type: number Default: 128

The size of the image when it is passed to be processed by the face detection model. The smaller the faster, but less precise in detecting small faces.

Must be divisible by 32, common sizes are: 128, 160, 224, 320, 416, 512, and 608.

scoreThreshold

Type: number Default: 0.5

Minimum confidence threshold for the face detection model.

Credits

• smartcrop.js
• face-api.js and the maintained fork vladmandic/face-api

License

MIT