phash-im

Perceptual image hashing via imagemagick.

There are many ways to compute perceptual image hashes, with the open source pHash library being one of the most popular. After researching a few dozen different approaches, we've settled on imagemagick's implementation, for the following reasons:

• modern - based on an approach from 2011
• robust against a variety of attacks
• thoroughly tested
• easy to install - imagemagick has great cross-platform support

Install

• Install a recent version of imagemagick >= v7 (brew install imagemagick on Mac OS).

npm install --save phash-im
# or 
yarn add phash-im

Usage

const phash = require('phash-im')
 
const hash1 = await phash.compute('./media/lena.png')
const hash2 = await phash.compute('./media/barbara.png')
 
const diff = await phash.compare(hash1, hash2)

API

phash.compute(input)

Returns: Promise<Array<Number>>

Computes the perceptual hash of the given image. Note that the result will be an array of 42 floating point values, corresponding to the 7 image moments comprised of 2 points each across 3 color channels RGB (7 * 2 * 3 = 42).

input

Type: String Required

Path to an image file.

phash.compare(hash1, hash2)

Returns: Promise<Number>

Computes the L2 norm of the two hashes returnd by phash.compute (sum of squared differences).

hash1

Type: Array<Number> Required

Perceptual hash of first image.

hash2

Type: Array<Number> Required

Perceptual hash of second image.

Related

• phash-imagemagick - Alternate version of this module which parses imagemagick output instead of using json.
• pHash - A popular open source perceptual hash library.

License

MIT © Travis Fischer