img-diff-js

:art: Node.js library to compare 2 images without native libs.

Actual	Expected	Difference

Install

npm install img-diff-js

const { imgDiff } = require('img-diff-js');

imgDiff({
  actualFilename: 'example/actual.png',
  expectedFilename: 'example/expected.png',
  diffFilename: 'example/diff.png',
}).then(result => console.log(result));

API Usage

imgDiff(opt: ImgDiffOptions): Promise<ImgDiffResult>

Create image differential between two images.

ImgDiffOptions

{
  actualFilename: string;
  expectedFilename: string;
  diffFilename?: string;
  generateOnlyDiffFile?: boolean; // default false
  options?: {
    threshold?: number;   // default 0.1
    includeAA?: boolean;  // default false
  }
}

• actualFilename - Required - Path to actual image file.
• expectedFilename - Required - Path to expected image file.
• diffFilename - Optional - Path to differential image file. If omitted, imgDiff does not output image file.
• generateOnlyDiffFile - Optional - Generate only files with difference
• options - Optional - An object to pass through pixelmatch.

ImgDiffResult

{
  width: number;
  height: number;
  imagesAreSame: boolean;
  diffCount: number;
}

• width - Differential image's width.
• height - Differential image's height.
• imagesAreSame - It'll be true only if 2 images are same perfectly.
• diffCount - The number of differential pixels.

Available format

The following codecs are available for input image files.

• png
• jpeg
• tiff (limited. See https://github.com/Quramy/decode-tiff#compatibility )
• bmp

imgDiff detects the input image format from it's extension name. For example, if the input file name ends with ".jpeg", imgDiff attempts to decode in JPEG way regardless of the actual file format.

The output image format is PNG only.

Performance

case name	img-diff-js	image-difference	image-diff
50 same dimension PNGs	936 msec	11018 msec	16539 msec
50 different dimension PNGs	715 msec	16077 msec	16639 msec
50 same dimension JPEGs	1076 msec	10910 msec	19078 msec

The above table was captured under Travis-CI. If you want the latest result, check the raw log.

Contribute

PR or issue is welcome :)

Setup

yarn

Test

yarn test

Run benchmark script

yarn run perf

License

MIT License. See LICENSE under this repository.