Humans can be tracked with unique 'fingerprint' based on how their bodies block Wi-Fi signals

Researchers detail WhoFi, a deep-learning approach that encodes Wi‑Fi Channel State Information to re‑identify individuals across locations, reporting up to 95.5% accuracy on the NTU‑Fi dataset using transformer-based models. By leveraging person-specific distortions in Wi‑Fi waveforms, the method enables non-visual, through‑wall tracking even without phones, illustrating expanding Wi‑Fi Sensing surveillance capabilities and associated privacy concerns.