Intel AI Introduces a Deepfake Detector that Analyzes “Blood Flow” in Video Pixels to Return Results in Milliseconds with 96% Accuracy

Deepfakes, the art of swapping one face for another, is probably fun. Technology has improved day by day and so have deepfakes and the biggest problem is knowing exactly how real deepfakes start to appear. Deepfakes are entertaining, but spoofing can spread false information, and we also have plenty of apps where some make fakes while others scramble to find them.

FakeCatcher, a new real-time deepfake detector from Intel, has an accuracy rate of 96% and can determine whether a video is fake or not. It’s the “world’s first real-time deepfake detector that delivers results in milliseconds,” according to Intel’s Responsible AI research. Most detectors work by looking at raw data, while FakeCatcher looks for genuine clues in real videos, like blood flow in video pixels, which means we’re human. Spatio-temporal maps are created by translating these signals, which are collected from all areas of the face. Researchers can then determine whether a video is genuine or fake using deep learning. This technology runs on a web-based platform that runs up to 72 different streams in parallel on 3rd Generation Intel Xeon Scalable processors.

Demir and Umur Ciftci of the State University of New York at Binghamton designed Intel’s FakeCatcher. This application uses Intel software and hardware and runs on a server and interface through a web-based platform. The FakeCatcher team used OpenVino to simulate and create AI models for the landmark and face detection algorithms. Computer Vision blocks have been optimized using OpenCV and Intel Integrated Performance Primitives. Inference blocks were optimized using Intel Advanced Vector Extensions 512 and Intel Deep Learning Boost, while media blocks were optimized using Intel Advanced Vector Extensions 2. The teams worked also used the Open Visual Cloud initiative to create an integrated software stack for the Intel Xeon Scalable processor. series.

The majority of deep learning-based detectors analyze raw data for indications of inauthenticity and identify flaws in a video. FakeCatcher, on the other hand, analyzes what makes us human in order to find real clues in real pictures, like our veins, which change color as our heart pumps blood. Blood flow signals are collected from the face and computers turn them into spatio-temporal maps. After that, we can quickly determine if a video is genuine or fake using deep learning.

Deepfake movies are a growing danger. It is difficult to detect these videos in real-time detection applications, as downloading and analyzing the videos usually takes hours. Deep deception can damage the reputation of the company and have negative impacts, such as lowering public trust in the media and by allowing users to distinguish between real and fake content, FakeCatcher promotes the trust. According to Gartner, companies will invest up to $188 billion in cybersecurity solutions.

For FakeCatcher, there are many potential applications. International news companies could use the detector to prevent the unintentional broadcast of doctored videos. Social media sites could use technology to prevent individuals from posting harmful deepfake videos. Additionally, charitable groups could use the platform to democratize deepfake detection for all users.

Check Inte reference articleI. All credit for this research goes to the researchers on this project. Also don’t forget to register. our Reddit page and discord channelwhere we share the latest AI research news, cool AI projects, and more.