Blockhash perceptual image hash algorithm

Latest on Hackage:

This package is not currently in any snapshots. If you're interested in using it, we recommend adding it to Stackage Nightly. Doing so will make builds more reliable, and allow to host generated Haddocks.

BSD-3-Clause licensed by Kwang Yul Seo

blockhash Build Status

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


Usage: blockhash [-q|--quick] [-b|--bits ARG] filenames

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


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)
comments powered byDisqus