Facial proximity detection based on inaudible noise for mobile authentication

Enterprise-class mobile devices find widespread use in the factories, warehouses, hospitals. A complement of audio communication features enables workers to maximize their productivity and improve customer service. We demonstrate that acoustic interaction of inaudible noise delivered by a mobile device with nearby physical objects has a potential for real-time proximity detection and facial recognition. The mobile device was mounted near the head and torso simulator inside the listening room. The device receiver produced low level inaudible equalized above 15 kHz noise signal, which was recorded by one of the device’s microphone. The received signal was measured at different device orientations and distances relative to the simulator’s artificial ear. The measurements revealed good correlation between received signal power spectral density and distance to artificial ear. Positive correlation with artificial ear shape was also observed. The results have applicability in the field of mobile communication security and potential for instant biometric authentication.