Skip to content

dannyi96/Image-Cryptography

Repository files navigation

Image Cryptography Based on Rubix's Cube Principle

Implementation of image encryption and decryption using Rubix's Cube Principle. This algorithm is based on the paper "A Secure Image Encryption Algorithm Based on Rubik's Cube Principle" by Khaled Loukhaoukha, Jean-Yves Chouinard and Abdellah Berdai.

Algorithm Overview

Given an input image having the three R,G,B matrices of size M X N Hyperparameters include α - used for vector creation ITER_MAX - maximum number of times to carry out operations

A. Encyption

  1. Create two vectors Kr and Kc with |Kr|=M & |Kc|=N. The values of these vectors are randomly picked from 0 to 2α -1

  2. Repeat below steps ITER_MAX number of times

    i. Rolling Rows:

    • The sum of all pixel values of every row of the image RGB matrices are calculated one by one.

    • If the sum of a given row rowNumber is even, Roll the row to the right Kr[rowNumber] times Otherwise roll to the left Kr[rowNumber] times.

    ii. Rolling Columns:

    • The sum of all pixel values of every column of the image RGB matrices are calculated one by one.

    • If the sum of a given row columnNumber is even, roll the column up Kc[columnNumber] times. Otherwise roll the column down Kc[columnNumber] times.

    iii. XORing Pixels:

    • For every pixel(i,j), XOR the pixel with the below two values

      • Value #1 - Kc[columnNumber] if i is odd else 180 rotated bit version of Kc[columnNumber]

      • Value #2 - Kr[rowNumber] if j is even else 180 rotated bit version of Kr[rowNumber]

B. Decryption

Given an encrypted image, vectors Kr and Kc & ITER_MAX , decryption can be done by following the reverse procedure - XORing pixels → Rolling Columns → Rolling Rows ITER_MAX number of times

Prerequisites

Running

  1. Using the crypto_client.py script supplying neccessary parameters
$ python3 crypto_client.py -h
usage: crypto_client.py [-h] [--type TYPE] [--image IMAGE] 
      [--alpha ALPHA] [--iter_max ITER_MAX] 
      [--key KEY] [--output_image OUTPUT_IMAGE]
  1. Using rubikencryptor python package
from rubikencryptor.rubikencryptor import RubikCubeCrypto
from PIL import Image

# Encrypt image
input_image = Image.open('image1.png')
encryptor = RubikCubeCrypto(input_image)
encrypted_image = encryptor.encrypt(alpha=8, iter_max=10, key_filename='key.txt')
encrypted_image.save('encrypted_image.png')

# Decrypt image
decryptor = RubikCubeCrypto(encrypted_image)
decrypted_image = decryptor.decrypt(key_filename='key.txt')
decrypted_image.save('decrypted_image.png')

Example

Original Image -

Run Encryption on the Original Image

$ python3 crypto_client.py --type encrypt
    --image example/original.png 
    --output_image example/encrypted.png 
    --key example/encoded_key.txt 
    --alpha 8 --iter_max 10

encrypted image is stored at example/encrypted.png & key is stored at example/encoded_key.txt

Encrypted Image -

Run Decryption on the Encryped Image using the Key

$ python3 crypto_client.py --type decrypt  
    --image example/encrypted.png 
    --output_image example/decrypted.png 
    --key example/encoded_key.txt

decrypted image is stored at example/decrypted.png

Decrypted Image -

Demo using a Web App ( Early beta version )

Note - this is still in early beta version. Performance enhancements, bug fixes & improved styling are expected.

App hosted at https://dannyi96.github.io/Image-Cryptography/

ui_phase1.mp4

Future of project

Refer the proposed Roadmap

Main goals include

  • Improving the Web App demonstrating this package
  • Measuring the current performance of the package ( encryption & decryption )
  • Improving the performance of the package ( encryption & decryption )
  • Improving Quality via Testcases, Linters, CICD pipeline
  • (Stretch goal) Including technical aspects of the research paper as to why the algo is resistant to various attacks as part of the WebApp