Skip to content

This project is a web application designed to detect deepfake videos using advanced machine learning techniques. The backend is powered by a pre-trained model that leverages Convolutional Neural Networks (CNNs) with InceptionV3 architecture and Recurrent Neural Networks (RNNs) utilizing Gated Recurrent Units (GRUs).

Notifications You must be signed in to change notification settings

Shiva-Khatter/DART---Deepfake-Analysis-and-Realism-Test

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

46 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DeepFake Video Detection

Introduction to DeepFakes

DeepFake is a combination of "Deep Learning" and "Fake," referring to the process of replacing someone's likeness in an image or video with someone else's using advanced technologies like Deep Artificial Neural Networks. This phenomenon has gained significant traction, particularly in the film industry, and poses a substantial threat to news organizations and the integrity of digital content. Major companies, including Google, are heavily investing in countering DeepFakes by releasing large datasets to train models that help combat this issue.

As Generative Adversarial Networks (GANs) that generate DeepFakes continue to improve, there is a pressing need for effective detection methods. The potential to create new GAN models that incorporate knowledge about combating existing models makes it increasingly challenging for current technologies to keep pace. This project aims to address this challenge by developing a robust tool for detecting deepfake videos.

Project Overview

This project is a web application designed to detect deepfake videos using advanced machine learning techniques. The backend is powered by a pre-trained model that leverages Convolutional Neural Networks (CNNs) with InceptionV3 architecture and Recurrent Neural Networks (RNNs) utilizing Gated Recurrent Units (GRUs).

Key Features

  • High Accuracy: Achieved an impressive accuracy in detecting deepfake videos.
  • Real-Time Analysis: Deployed using Flask, enabling real-time video analysis.
  • Scalability: Successfully processed about 450 video uploads during testing, demonstrating the tool's efficiency and reliability.

Deployment

This application can be easily run on your local machine or deployed to the cloud, making it accessible for various use cases.

Initially, the focus was on simply detecting forged faces, followed by the development of a comprehensive model for detecting deepfake videos.

I am willing to share the code, dataset and provide further information if needed. Feel free to reach out with any questions or inquiries!

For accessing the dataset kindly seek access and share a note on what kind of project you are doing and require the dataset for!

Feel free to reach out!!

About

This project is a web application designed to detect deepfake videos using advanced machine learning techniques. The backend is powered by a pre-trained model that leverages Convolutional Neural Networks (CNNs) with InceptionV3 architecture and Recurrent Neural Networks (RNNs) utilizing Gated Recurrent Units (GRUs).

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published