dart_imagehash

A Dart package for perceptual image hashing, inspired by the Python imagehash library.

Image hashing algorithms generate compact, fixed-length fingerprints from images that allow you to:

Find visually similar images (even with small transformations)
Detect duplicate or near-duplicate images
Perform content-based image retrieval

Features

This package provides multiple image hashing algorithms:

Average Hash (aHash)
Perceptual Hash (pHash)
Difference Hash (dHash)
Wavelet Hash (wHash)

Getting started

Add this to your pubspec.yaml:

dependencies:
  imagehash: latest_version

Then run:

dart pub get

Usage

import 'dart:io';
import 'package:image/image.dart' as img;
import 'package:dart_imagehash/dart_imagehash.dart';

void main() {
  // Load images
  final image1 = img.decodeImage(File('image1.jpg').readAsBytesSync())!;
  final image2 = img.decodeImage(File('image2.jpg').readAsBytesSync())!;

  // Calculate hashes
  final hash1 = averageHash(image1);
  final hash2 = averageHash(image2);

  // Calculate the similarity (0-1, where 1 is identical)
  final hashDistance = hash1 - hash2;
  final similarity = 1.0 - (hashDistance / (hash1.hashSize * hash1.hashSize));

  print('Hash 1: ${hash1}');
  print('Hash 2: ${hash2}');
  print('Distance: $hashDistance');
  print('Similarity: ${(similarity * 100).toStringAsFixed(2)}%');

  // Try other hashing algorithms
  final pHash1 = perceptualHash(image1);
  final pHash2 = perceptualHash(image2);
  print('Perceptual Hash distance: ${pHash1 - pHash2}');
}

For a more comprehensive example that demonstrates comparing images with all four hash algorithms, check out the example included in this package. The example demonstrates how to:

Calculate and compare hashes for similar and different images
Display similarity percentages for each algorithm
Handle file loading with relative paths

Hash Comparison

Image hashes are compared using the Hamming distance - the number of bits that differ between two hashes. A smaller distance indicates more similar images.

// Get the Hamming distance
int distance = hash1 - hash2;

// Boolean comparison (only true if exactly equal)
bool identical = hash1 == hash2;

// Convert hash to hex string
String hexString = hash1.toString(); // or hash1.toHex()

// Create hash from hex string
var hash = ImageHash.fromHex('f8e0a060c020f8e0', 8);

Additional information

This package is a Dart implementation of the algorithms found in the Python imagehash library. For more information about the theory behind perceptual hashing, see: