Quick, simple implementation of Error Level Analysis

ela.py

#!/usr/bin/env python

# This is a really simple implementation of ELA as described in
# http://blackhat.com/presentations/bh-dc-08/Krawetz/Whitepaper/bh-dc-08-krawetz-WP.pdf
# You shouldn't actually use it, or at least read the paper carefully
# and implement more of the techniques before drawing any conclusions.

from PIL import Image, ImageChops, ImageEnhance
import sys, os.path

filename = sys.argv[1]
resaved = filename + '.resaved.jpg'
ela = filename + '.ela.png'

im = Image.open(filename)

im.save(resaved, 'JPEG', quality=95)
resaved_im = Image.open(resaved)

ela_im = ImageChops.difference(im, resaved_im)
extrema = ela_im.getextrema()
max_diff = max([ex[1] for ex in extrema])
scale = 255.0/max_diff

ela_im = ImageEnhance.Brightness(ela_im).enhance(scale)

print "Maximum difference was %d" % (max_diff)
ela_im.save(ela)
ela_im.show()