Dopious
Senior Member
Founding Member
Sapphire Member
Patron





Up to 95 percent accuracy.
A group of researchers at La Sapienza University in Italy has shown that Wifi signals can be used to physically identify and track people with high accuracy.
In an article published in mid-July, the researchers review how they can produce "fingerprints" based on how Wifi signals are reflected and refracted by bodies, using machine learning and artificial neural networks.
Experiments have shown that these “fingerprints” are stable enough that people can be recognized even if they change clothes or are wearing a backpack. The accuracy is up to 95 percent (average recognition 88 percent), which according to the researchers is in line with other methods for automatically recognizing individuals in different contexts.
The researchers call their technology “WhoFi” and present it as a possible solution to the problem of re-identifying people who have been seen in one location via surveillance camera and then found again in the feeds from surveillance cameras in other locations.
Source: https://arxiv.org/html/2507.12869v1
A group of researchers at La Sapienza University in Italy has shown that Wifi signals can be used to physically identify and track people with high accuracy.
In an article published in mid-July, the researchers review how they can produce "fingerprints" based on how Wifi signals are reflected and refracted by bodies, using machine learning and artificial neural networks.
Experiments have shown that these “fingerprints” are stable enough that people can be recognized even if they change clothes or are wearing a backpack. The accuracy is up to 95 percent (average recognition 88 percent), which according to the researchers is in line with other methods for automatically recognizing individuals in different contexts.
The researchers call their technology “WhoFi” and present it as a possible solution to the problem of re-identifying people who have been seen in one location via surveillance camera and then found again in the feeds from surveillance cameras in other locations.
Source: https://arxiv.org/html/2507.12869v1