Skip to content

ManuFU/Detector-for-fake-faces-

Repository files navigation

Fake Face Detector

The goal of the detector is to distinguish real faces from generated ones. For this purpose thousands of photos of people (https://github.com/NVlabs/ffhq-dataset) and thousands of pictures of "generated" people (https://futurism.com/incredibly-realistic-faces-generated-neural-network) were used. The detector was trained with the library FastAI and the technique Transfer Learning. Accuracy was surprisingly good, with over 90% on test data. Further action in the area of Explainable AI is needed to verify that the model is robust.

Login

- Login in with Username: admin Password: admin

Forward to payment option

- Click on "Pay with Card"

Payment (Stripe)

- Fill the boxes
- Use the credit card number 4242424242424242

Prediction / Inference

- Upload a photo with a face on it - Click on "Predict"

Setup

  • Clone or download repo
  • Insert your Stripe API key in app.py
  • Deploy at render.com or somewhere else

Boilerplate

git clone https://github.com/anfederico/Flaskex
cd Flaskex
pip install -r requirements.txt
python app.py



Features

  • Encrypted user authorizaton
  • Database initialization
  • New user signup
  • User login/logout
  • User settings
  • Modern user interface
  • Limited custom css/js
  • Easily customizable



About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published