Voice Processing by Wideband Accelerometers with Immunity to Environmental Acoustic Noise
Enrico Alessi, Ivana Guarneri, Fabio Passaniti, Michele Dellutri · 2022 IEEE Sensors · 2022
Voice is the key natural communication channel for humans. Machines and consumer devices are reshaping themselves for interacting with humans by voice. In this domain artificial intelligence (AI) and machine learning techniques are extensively used for cleaning up the microphone-based voice signal from the environmental acoustic noise and other artifacts. The presented work is focusing on an alternative or complementary approach to microphones for voice data capturing and processing based on the latest generation of motion sensors. A vocal command classifier of bone conduction voice signals is presented for consumer market and for impaired speech people as well.