インテル、世界初のリアルタイムディープフェイク検出器を発表

Intel on November 14 introduced what it claims to be the world’s first deepfake detector that can return results in milliseconds with 96% accuracy rate. Namely, FakeCatcher which analyzes blood flow of the human subject in pixels of a video through photoplethysmography (PPG).

Intel Deepfake Detector Ci

“Most deep learning-based detectors look at raw data to try to find signs of inauthenticity and identify what is wrong with a video. In contrast, FakeCatcher looks for authentic clues in real videos, by assessing what makes us human – subtle “blood flow” in the pixels of a video,” Intel explained in a blog post.

Deepfakes are usually fake videos that use artificial intelligence to copy a person’s face to another in an attempt to convince people that they are the one who appears on the video. In contrast to real videos, the blood flow may be absent in the deepfakes.

“When our hearts pump blood, our veins change color. These blood flow signals are collected from all over the face and algorithms translate these signals into spatiotemporal maps. Then, using deep learning, we can instantly detect whether a video is real or fake,” Intel said.

According to Intel, the technology’s potential use cases include social media platforms to prevent their users in uploading harmful deepfake videos; global news organizations to avoid amplifying manipulated videos; and non-profit organizations “to democratize detection of deepfakes for everyone.”

Intel reasoned out that deepfake videos are a growing threat and cited research firm Gartner that companies spend up to USD 188 billion in cybercrime solutions. FakeCatcher’s ability to detect whether a video is fake or not in real-time immediately prevents further spread of deepfake that might influence people. This is something that other existing deepfake detection apps lack that usually require uploading videos for analysis, thus waiting hours for results.

“Deception due to deepfakes can cause harm and result in negative consequences, like diminished trust in media. FakeCatcher helps restore trust by enabling users to distinguish between real and fake content,” Intel said.

Source (1)

What is Intel's FakeCatcher and how fast is it?

FakeCatcher is a real‑time deepfake detector that returns results in milliseconds with a 96% accuracy rate.

How does FakeCatcher determine whether a video is real or fake?

The system analyzes subtle blood‑flow signals in facial pixels, creates spatiotemporal maps, and applies deep learning to classify authenticity.

Which sectors does Intel envision using this technology?

Potential users include social‑media platforms, global news organizations, and non‑profit groups aiming to democratize deepfake detection.

React to this article: