This repository contains the project notebook that was developed during the Module #3 Deep learning, computer vision and evolutionary computation course by TDX | University of Coimbra Data Science Academy (Powered by Talkdesk). Total course time was 80 hours plus 3 additional weeks for project related work.
For the project work and with the provided dataset, a real-fake face classifier was developed using several approaches: convolutional neural networks (CNN) and transfer learning with and without face cropping using the Multi-Task Cascaded Convolutional Neural Networks (MTCNN).
To extend the course work, and in order to leverage the created models, a web application was developed using the streamlit framework. For more information about the models or the project, check out the project folder in the root directory of this repository. The application source code is in the app folder.
Open a terminal and start by cloning the project repository
git clone https://github.com/dvpinho/RealFakeFaces.gitGo to the project directory
cd RealFakeFacesInstall dependencies
pip install -r requirements.txtStart the server (Linux / MacOS)
streamlit run app/app.pyStart the server (Windows)
streamlit run app\app.pyIf the browser window does not open automatically when the streamlit run command is executed, you can manually navigate to localhost:8501
To terminate the application, go back to the terminal and shutdown the server by pressing CTRL + C.
