This is a perceptual image hash calculation tool based on algorithm described in Block Mean Value Based Image Perceptual Hashing by Bian Yang, Fan Gu and Xiamu Niu. Visit the website for further information.

Installation

From Hackage

cabal install blockhash

From source

stack build

Program

Usage: blockhash [-q|--quick] [-b|--bits ARG] filenames
  blockhash
Available options:
  -h,--help                Show this help text
  -q,--quick               Use quick hashing method
  -b,--bits ARG            Create hash of size N^2 bits.

Library

The library exposes the Data.Blockhash module with the following API:

blockhash - Calculate perceptual hash for an RGBA image
hammingDistance - Calculate the hamming distance between two hashes
Image - Image data type (width, height, RGBA pixels)
Hash - Hash result type
Method - Hashing method (Precise or Quick)

The example code below uses JuicyPixels to load images and prints the hash to stdout.

import qualified Codec.Picture as P
import Data.Blockhash
import qualified Data.Vector.Generic as VG
import qualified Data.Vector.Unboxed as V
printHash :: FilePath -> IO ()
printHash filename = do
  res <- P.readImage filename
  case res of
    Left err -> putStrLn ("Fail to read: " ++ filename)
    Right dynamicImage -> do
      let rgbaImage = P.convertRGBA8 dynamicImage
          pixels = VG.convert (P.imageData rgbaImage)
          image = Image { imagePixels = pixels
                        , imageWidth = P.imageWidth rgbaImage
                        , imageHeight = P.imageHeight rgbaImage }
          hash = blockhash image 16 Precise
      putStrLn (show hash)