Hidden people could be identified using Wi-Fi signals
Image credit: E&T editorial staff
Researchers at the University of California-Santa Barbara have presented a new method which allows for a pair of Wi-Fi transceivers to be used to detect if a person behind a wall is the same person featured in a video.
The model uses the unique gait of the individual to confirm whether they are the same person featured in video footage. Gait identification involves tracking the movement of each joint in the body to compare to a reference or to determine characteristics such as height, age and sex. This technique has the advantage of being possible to do at a distance, unlike types of biometric identification which rely on irises, faces or fingerprints.
While video footage is typically considered the minimum requirement to perform gait identification, this new method requires only a pair of Wi-Fi transceivers.
“Our proposed approach makes it possible to determine if the person behind the wall is the same as the one in video footage, using only a pair of off-the-shelf Wi-Fi transceivers outside,” said Professor Yasamin Mostofi, who led the study. “This approach utilises only received power measurements of a Wi-Fi link. It does not need any prior Wi-Fi or video training data of the person to be identified. It also does not need any knowledge of the operation area.”
Mostofi and her team experimented with placing one Wi-Fi transmitter and one Wi-Fi receiver behind walls, outside a room in which a person is walking. The transmitter sends a signal and the signal’s power is measured by the receiver; these measurements are used to compare it with the gait of a person captured in a video. However, this required the researchers to take the video footage and transform it into something which could be meaningfully compared to the data collected from the hidden person.
The researchers used an algorithm to extract a mesh from the video footage, which describes the surface of the body as time passes. They then simulated the electromagnetic signal that would be generated if this person walked through a Wi-Fi area. Key gait features were extracted from both the simulated and measured signal, allowing for them to be compared to determine if the hidden person is the same as the person in the video.
The researchers have “extensively” tested the method on campus with 1,488 Wi-Fi video pairs, drawn from a pool of eight people in hidden in three different areas. The technique could be used to correctly identify the hidden person 84 per cent of the time.
Mostofi and her colleagues hope that the system could be used in a variety of applications, such as surveillance, security and in smart homes. For example, law enforcement could determine whether a person hiding in a house is the same person who committed an offence captured in CCTV footage.
Sign up to the E&T News e-mail to get great stories like this delivered to your inbox every day.