Intel launched a FakeCatcher testing tool to determine whether it is an AI face change by detecting facial blood vessels.

Thanks to CTOnews.com netizen OC_Formula for the clue delivery! CTOnews.com, November 15 (Xinhua) Intel has developed a Deepfake detection tool called FakeCatcher, which claims to have 96% accuracy in detecting videos tampered with by AI. More importantly, Intel says FakeCatcher is a real-time test that produces results in milliseconds.

Deepfake videos are fake videos made using artificial intelligence algorithms that can attach one person's face to another's face or even imitate their voice. these techniques are very risky to be used to do evil.

Not only does Deepfake spread the wrong message that nothing actually happened, criminals can also claim that the video was made by Deepfake to get off the hook, and there is a great need for a system that can distinguish between fact and fiction.

Intel's FakeCatcher is used to detect Deepfake, and its principle is to judge by detecting blood flow. When blood flows through our bodies, our blood vessels change color in very subtle ways that are difficult for the human eye to detect, but can be detected by optical densitometry (PPG).

FakeCatcher detects these subtle PPG signals in video clips and converts them into space-time maps, meaning that they detect where changes occur and how they change over time. Intel's artificial intelligence engine can then compare these maps with its training set and mark them immediately if anything goes wrong.

CTOnews.com learned that the Intel lab team powered FakeCatcher with Intel's third-generation Xeon scalable processor, which has up to 72 concurrent detection streams and operates through a web interface.

96% real-time accuracy is really excellent, and existing solutions, including Adobe solutions, are generally carried out asynchronously. The video must be sent to the server for processing and it will take several hours to make a judgment.