Skip to content

Classifies a given aadhaar image to real or fake by doing two levels of analysis.

Notifications You must be signed in to change notification settings

z1311/Fake-Aadhaar-Detection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Fake Aadhaar Detection

This project classifies the scanned aadhaar image to either real or fake image by doing two levels of testing. Fake images are the images that are digitally altered.

Level 1 Testing

Whenever an image is altered using software tools they leave software signatures in the metadata of the image. Level 1 testing exploit this feature and tries to findout traces of any signature. It is the fastest and simplest way to classify but there are online tools/websites that helps to clean this type of information in metadata. MS Paint is a good example that doesn't attach its signature in the metadata of the image.

Level 2 Testing

It starts of by doing ELA (Error Level Analysis) on the image and the result is given to LBPH (Local Binary Patterns Histograms) recognizer which decides whether it is fake or real image. Even though LBPH algorithm is used for face recognition, it can be used in this project for generating histograms and comparing them.

Block Diagram

Untitled Diagram

Output Screenshots

For selecting image
Capture1
Fake image output screen
Capture2 Capture6
Real image output screen
Capture4 Capture5

Tools Used

  1. Python 2.7.14
  2. OpenCV2
  3. Tkinter
  4. Pillow
  5. Numpy