Skip to content

HackBMU/HackBMU2020_Luciole

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 

Repository files navigation

hackBMU2020_Luciole

Real vs Fake

Rumors. Gossip. Fake news. We've all heard these terms, messaging app like Facebook and Whatsapp has become a major channel for hate speech and false stories in India. The government is demanding changes. A simple way to identify the fake images by appling CNN on more than 10M fake and original images using tenserflow. Our project aims to detect if any changes are made in the images after it was clicked.

Today rumors can easily spread on the internet as people are exponentially increasing and with such low internet prices everyone is eager to go online. Our project can help in detecting the fake images and can tell if the image is original or doctored.

This can help society as people will have the right information and will not believe in doctored images.

Challenges we ran into:

The data was easy to access but it was in bits and peases. So, preparing the data took a lot of time. After passing it through a neural network, the accuracy was about 56% the 1st time. It even falls every time we tried to make it better. This was the hardest task for us in this project but finally, it has an accuracy of about 99%. We are using the dataset "Real and Fake Face Detection" which contains real/fake face photos, respectively. We have implemented our project in python with the help of following libraries:

  1. opencv2
  2. matplotlib
  3. keras
  4. numpy
  5. pandas

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages