auto-forgery-detection

Automated Image Forgery Detection through Classification of JPEG Ghosts

Template

Fast start.

Try code below and you will see the magic:

from manage import *
path = 'data/trump-queen-fake.jpg'
predict_img(path, w=16)

How to

How to start

1. In console git clone https://github.com/kalinkinisaac/auto-forgery-detection.git
2. Set up your virtual enviroment (python3)
3. Install packages pip install -r requirements.txt
4. Go to the folder cd auto-forgery-detection/
5. You are ready to go! You are perfect!

How to check photo

1. Open python console, when you are in project folder python
2. Import all from managy.py from manage import *
3. Let be you photo you want to analise be path = /Users/user/Downloads/test.jpg
4. Determine path to photo in console path = 'path'
5. Run prediction script predict_img(path, w=64)
6. You will see your image, and the prediction process will start.
7. If your image analisys is not accurate, than you should decrease value of w. For example here you can see difference w=64 to the right, and w=16 to the left:
8. After you guessed the right w=w0, than you are able to predict finally.

Green and Red squares (named blocks) represents difference between to images.

Used sources

• Article: 'Automated Image Forgery Detection through Classification'
• Documentation of scikit-learn 0.21.2
• NumPy User Guide

Coded and commited by

• Kalinkin Isaac (@kalinkinisaac)
• Nikolenko Alex (@GorillazZz1)
• Novik Ivaan