Deepfake Prevention

  • Home
  • Deepfake Prevention

How to detect Deepfakes

divider 1

What are Deepfakes?

divider

Deepfakes are computer generated videos of individuals in which one person can impersonate someone else’s likeness. They involve of machine learning techniques such as generative adversarial networks (GANs).

Danger sign

Why are they a problem?

divider

Deepfakes can be used for identity fraud, causing huge problems for the biometric authentication industry. For example, the technique can be used in a video spoof attack whereby a deepfake is matched with a stolen ID to bypass authentication checks. This is causing a growing concern for biometric authentication firms due to the increase in the frequency and success rate of fraudulent attacks, leading to lost revenue and consumer confidence.

How we solve this

divider

At Vastmindz we are now able to provide an enhanced biometric authentication check by verifying that a person in a video exhibits human physiology. 

We generate a spectrogram (see image) of their physiological signal to determine this. A real person’s spectrogram will have a clear singular beam, but deepfakes will not.

Our process

divider

Facial scan

Reading will last between 10-15 seconds.

Analysis in progress

Our solution will detect for human physiology.

Confidence score

Output probability value showing the likelihood of authenticity.

Happiness

Outcomes you can expect

divider
  • Verified human authentication
  • Increased validation speed
  • Reduced fraud
  • Improved customer experience 

AI-powered insights for
health & wellness

Our solutions are not a replacement for a healthcare professional and our solution does not diagnose, prevent, provide any treatment for any form of illness or disease.

Our Solutions

Rectangle 412
Sign up to get the latest on sales, new releases and more…

Copyright 2022 Vastmindz AI Ltd All Rights Reserved.