Deepfakes Detection in Germany through Images

Local Chapter Münich, Germany Chapter

Coordinated byGermany ,

Status: Completed

Project Duration: 27 Feb 2023 - 17 Mar 2023

Open Source resources available from this project

Project background.

Deepfakes are artificial intelligence generated or manipulated multimedia content (images, videos) that depict real people doing or saying things they never did in reality. Due to the recent advancements in architectures like Generative Adversarial Networks (GANs), deepfake generation became considerably faster and at a lower cost, presenting a significant threat to information media, politics and cyber security through spreading false information, identity theft and manipulating public opinion. Hence the necessity of having effective deepfake detection mechanisms.

The problem.

Deepfake generators have been able to produce imitations almost undetectable through human inspection. Per se, deepfake detection is one of the notable challenges of digital forensics and media security. They can be used to:

– Spread false information and manipulate public opinion.
– Defame, harass or blackmail individuals.
– Damage the credibility of media information and undermine trust.
– Spread fake news, political propaganda, and manipulate elections.

An artificial intelligence (AI) solution can help identify deepfake images with accuracy and precision.

Project goals.

Develop a deep learning model able to detect forgeries in images.

Project plan.

  • Week 1

    Research previous work and Data Collection

  • Week 2

    Exploratory Data Analysis

  • Week 3

    Preprocessing and Augmentation

  • Week 4

    Model Development

  • Week 5

    Model Training

  • Week 6

    Model Analysis and Interpretation

Learning outcomes.

Deep Learning, GAN, CNN, team work, project management.

Share project on: