Published a library that hides character data in Python images

Published on 2016.04.01 A library that hides character data in images using Python. You can also run it from the command line. Works with Python2 series.

install

install

pip install steganography

• pypi: steganography

Data embedding example

Sample code

sample-code

# -*- coding: utf-8 -*-
from __future__ import absolute_import, unicode_literals
from steganography.steganography import Steganography

#Embed text in image
path = "/tmp/image/a.jpg "
output_path = "/tmp/image/b.jpg "
text = 'The quick brown fox jumps over the lazy dog.'
Steganography.encode(path, output_path, text)

#Load the text hidden in the image
secret_text = Steganography.decode(output_path)

Sample command

python

#Embed text in image
>>>steganography -e /tmp/image/a.jpg /tmp/image/b.jpg 'The quick brown fox jumps over the lazy dog.'

#Load the text hidden in the image
>>>steganography -d /tmp/image/b.jpg
The quick brown fox jumps over the lazy dog.

1. Operation overview-text embedding

1-1. normalize Scan all the pixels in the image and select the pixels that have a remainder of 1 when each RGB value is divided by 8. Change the color of the selected pixel so that the remainder is not 1.

■ Pixels whose color has been changed by normalize processing

1-2. Text hiding

Convert text to hexadecimal. Divide the image into 16 pixels and change the value of the pixel color of the corresponding part so that the remainder becomes 1 when each RGB value is divided by 8.

2. Operation overview-Read text hidden in the image

2-1. Read text

Read the image with the text hidden, divide the image into 16 pixels, and identify the point pixels where the pixel color is 1 when each RGB value is divided by 8. The hexadecimal value is acquired from the point pixel position, converted from the hexadecimal number to a character string, and output.

3. Sample implementation code for study

3-1. Code that opens an image and reads it pixel by pixel

Images can be easily manipulated by using pillow.

open-image

# -*- coding: utf-8 -*-
from __future__ import absolute_import, unicode_literals
from PIL import Image

path = "/tmp/image/a.jpg "
img = Image.open(path)
img = img.convert('RGB')

for y in range(img.size[1]):
    for x in range(img.size[0]):
        r, g, b = img.getpixel((x, y))
        print(r, g, b)

output

>>> python sample.py
(44, 81, 110)
(75, 109, 137)
(85, 114, 144)
(69, 99, 127)
(68, 105, 131)
(52, 94, 116)
(65, 103, 124)
(108, 130, 154)
(149, 150, 181)
(164, 148, 184)
(152, 132, 169)
....

3-2. Code to change the pixel color of a specific condition to red

red

# -*- coding: utf-8 -*-
from __future__ import absolute_import, unicode_literals
from PIL import Image

path = "/tmp/image/a.jpg "
output_path = "/tmp/image/sample_red.png "
img = Image.open(path)
img = img.convert('RGB')


def matched_pixel(r, g, b):
    return r % 8 == g % 8 == b % 8 == 1


for y in range(img.size[1]):
    for x in range(img.size[0]):
        r, g, b = img.getpixel((x, y))
        if matched_pixel(r, g, b):
            #Rewrite the specified coordinates to red
            img.putpixel((x, y), (255, 0, 0))

img.save(output_path, "PNG", optimize=True)

3-3. str hex converter

str-hex-converter

# -*- coding: utf-8 -*-
from __future__ import absolute_import, unicode_literals


def to_hex(s):
    return s.encode("hex")


def to_str(s):
    return s.decode("hex")


base = "aiueo kakikukeko"
print(to_hex(base))
# >>> 616975656f206b616b696b756b656b6f

assert base == to_str(to_hex(base))

4. Source code

Pillow is used to read and write images. The source code of the library is available on github. https://github.com/subc/steganography/blob/master/steganography/steganography.py