The researchers based at the Indian Institute of Technology (IIT) – Ropar in Punjab (India) and Monash University in Melbourne (Australia) have developed a unique detector called the FakeBuster.
The members on the developing team include Associate Professor Ramanathan Subramanian, Dr Abhinav Dhall, Vineet Mehta and Parul Gupta.
According to the research team, the new device can be attached to laptops and desktops:
“FakeBuster is a standalone deep learning based solution, which enables a user to detect if another person’s video is manipulated or spoofed during a video conferencing based meeting.”
This new software could prove to be a major step towards exposing imposters who attend virtual conferences without the knowledge of the organiser.
The FakeBuster will expose imposters and also detect faces manipulated on social media to defame someone.
In a statement, Dr Abhinav Dhall said that sophisticated AI techniques have spurred a dramatic increase in the manipulation of media contents and they keep evolving and becoming more realistic.
“The tool has achieved over 90 per cent accuracy.”
It will also detect if an individual is attending a meeting on behalf of a colleague by morphing his image with his own.
A paper titled “FakeBuster: A DeepFakes Detection Tool for Video Conferencing Scenarios” was presented at the 26th International Conference on Intelligent User Interfaces in the US in April 2021.
The software, researchers say, has been tested with Zoom and Skype.
According to its developers, the FakeBuster tool works in both online and offline modes.
“This tool is independent of video conferencing solutions and has been tested with Zoom and Skype applications. It uses a 3D convolutional neural network for predicting video segment-wise fakeness scores.”
The team asserts that the FakeBuster is one of the first tools to detect imposters during video conferencing using DeepFake detection technology.
This tool developed by the Indian and Australian research team is expected to hit the market soon.
WATCH VIDEO: FakeBuster Intro
