▶ Comprehensively detect video, audio, etc. to determine authenticity within 5 to 10 minutes.. Modulation rate and synthesis types also available
▶ Optimized for use by domestic companies and institutions, training a total of 5.2 million data, including 1 million Korean
▶ Plans to build solutions in various fields... Anyone can purchase through the official website
DeepBrain AI (CEO Se-young Jang), a generative artificial intelligence (AI) company, announced on the 00th that DeepBrain AI has collaborated with the National Police Agency to introduce a deepfake detection solution to help crack down on deepfake crimes.
As concerns about crimes abusing deepfake technology have increased due to the recent advancement of deepfake technology, Deepbrain AI, cooperate with the National Police Agency, has developed a customized deepfake detection solution that can effectively respond to. The National Police Agency plans to use this technology to actively respond to deepfake-based intelligent crimes such as election crimes targeting the general election, phishing crimes, and synthetic sexual exploitation crimes.
DeepBrain AI's deepfake detection solution consists of ▲comprehensive detection ▲voice detection supports image, video, and voice detection respectively. If user uploads a video suspected of being a deepfake to the system and set the ▲ detection model, ▲detection period, and ▲person to be detected, its authenticity is immediately determined.
Images and videos are analyzed on a pixel-by-pixel basis, and audio is detected by comprehensively considering various operating factors such as frequency, time, and noise. Detection time usually takes about 5 to 10 minutes, and if it’s found to be a deepfake, it will be marked as fake. Once the determination is complete, technical information used in the crime, such as modulation rate and synthesis type, is also provided.
DeepBrain AI’s Deepfake detection solution boasts excellent completeness as it is designed based on generative AI avatar solution and a large amount of internally accumulated deep learning-based AI human data. Unlike most existing detection models that were created with data based on Westerners, DeepBrain AI model’s detection rate was increased by training from a total of 5.2 million of data, including 1 million of Korean data and 130,000 of Asian racial data.
DeepBrain AI plans to use this collaboration with the National Police Agency as an opportunity to discuss ways to build a deepfake detection solution with companies and organizations in various fields; plans to actively provide solutions that immediately determine the authenticity of false information that causes confusion and crisis situations and minimize the cost of loss to society.
Anyone can purchase and use Deepbrain AI's deepfake detection solution by making an inquiry on the official website (www.deepbrain.io). It is provided based on software as a service (SaaS), be easily used anytime, anywhere without the need to build separate servers and equipment.
Jang Se-young, CEO of Deepbrain AI, said, “In order to effectively respond to new deepfake crimes, we will continue to enhance our deepfake detection solution in cooperation with the National Police Agency”, “The ultimate goal is to neutralize deepfake crimes by helping anyone easily and conveniently judge false facts,” he said. (end)
