WiFi Signals Used to Generate Human Images and Identify Individuals

A study reveals that unencrypted WiFi signals exchanged between devices and routers can be analyzed to create radio-based images of people and recognize them with high accuracy. Researchers demonstrated that by observing how radio waves propagate through spaces, they can generate detailed images of individuals and their surroundings. In tests involving 197 participants, the system achieved nearly 100% identification accuracy regardless of movement or angle. Professor Thorsten Strufe from KIT’s Institute of Information Security explained that this technique works similarly to a camera but uses radio waves instead of light. The method leverages beamforming feedback information (BFI) — unencrypted data sent by devices to optimize wireless communication — to map human shapes and detect presence. Once trained, machine learning models can identify individuals in seconds without requiring the target to carry WiFi-enabled devices. This raises significant privacy concerns, as passive observation of ambient signals could enable surveillance without direct interaction. Commenters suggested countermeasures like wired connections or tin foil suits, while others noted the technique’s similarity to radar systems. The research highlights both potential applications in security and the risks of pervasive data collection through everyday technology.