Researchers at the University of California, Santa Barbara, have invented a new way that allows them to identify people behind a wall, using wireless WiFi transceivers and analyzing their gait.
This new method, called XModal-ID (pronounced Cross-Modal-ID), was developed by Yasamin Mostofi and her team and offers the ability to identify a person behind a wall if the same person has previously been captured on video. And all this is done with the exclusive use of WiFi signals.
How does it work?
First, a 3D image of a person must be extracted from video, using a Human 3D Mesh Extraction algorithm
The researchers then try to simulate the radio frequency signals that would have been generated if the person were walking in an area covered by WiFi, through an approximation of electromagnetic waves.
They then try to match the face's gait by comparing spectrograms created from real-time WiFi signals and the video in which it has been recorded.
According to the researchers, in the paper they published, the technique does not require training a system in facial recognition or knowledge of the environment. The process only needs regular WiFi transceivers to work.
The researchers ' tests showed an accuracy of between 82% and 89%, making the technique quite effective. The team says it will particularly help authorities identify suspects when investigating a crime.
There have been other tests in the past
This isn't the first time researchers have developed a way to recognize people through walls. In 2015 MITdeveloped an RF receiver that could recognize human silhouettes behind a wall using reflections from the human body.
Researchers at the Technical University of Munich (TUM) also developed a method for creating 3D images of the surrounding space using Wi-Fi signals in 2017. And last year, a team of researchers from MIT developed an AI for the same purpose.
It is clear that researchers around the world are trying to harness the power of WiFi to identify people behind walls. At the same time, there are many privacy concerns raised by such techniques, since almost everyone has a WiFi router at home.
