Intel Detection Tool Uses Blood Flow to Identify Deepfakes with 96% Accuracy

“Deepfake videos are everywhere now. You have probably already seen them; videos of celebrities doing or saying things they never actually did.”

–Ilke Demir, senior staff research scientist in Intel Labs

What is a deepfake?

A deepfake is typically superimposes someone's face and voice onto another person in real time, this first gained traction when adult websites began banning videos where the technique was used to add famous actresses' faces to porn stars bodies.

Image of a wooden clothing model meant to look human, infant of garments.

The science behind it:

There is a new tool now to identify deepfake videos, this includes using the science behind the fact that our veins change color when blood is pumped through them. This is called photoplethysmology or PPG used in mapping. Intel developed this software for FakeCatcher their new program to identify deepfakes. The last piece of the puzzle to solving deepfake problems is using eyegaze detection platforms. We also know now that deepfakes use different faces over the same facial expressions so if we can pick up on facial expressions that have commonalities, we can identify the deepfake with an almost 96% accuracy rating.

What is the need for deepfake detection platforms?

There are many nefarious purposes that deepfakes are used for, including fake news, pornography, and mainly political purposes. There are positives to this new software such as using deepfakes to render younger versions of actors in films or recreate actors who have passed away in movies. However, past the world of cinematography and within it, people should have the right to identify whether something is a deepfake or not. Thus deepfake detectors come in handy.

Protestors carrying sign, the main one in focus stating “I wish this were fake news”"

Another necessity this technology is being used for is finding scammers applying for remote jobs as well as an appdesigned to remove women's clothing digitally.

The biggest concern being how deepfakes have led to the spread of misinformation in the fake video of Ukraine president surrendering was spread on social media earlier this year.

Intel states "Real-time deepfake detection empowers people, governments, and the media to instantly identify what's real." from their video on what makes someone human.

Intel is not alone in this fight to identify deepfakes, Facebook, the Defense department, Adobe and Google have all been working on creating the same tools to identify deepfakes. Intel is simply the first one to be analyzing blood flow signals.

The future of deepfakes

This new technology could be very useful but it could also make the technology that creates these fake people even better, I guess we'll have to wait and find out. We'll be updating our audience here as the story progresses.

Previous
Previous

How much privacy does your iPhone really give you? A Deepdive

Next
Next

Predicted 40% Labor Drop in Electric Vehicle Industry