SEOUL -- Inspired by basilar membrane inside the cochlea, a spiral-shaped cavity inside human ears, South Korean researchers claimed to have developed a self-powered flexible sound sensor which can be used by artificial intelligence to identify people by their voices.
The Korea Advanced Institute of Science & Technology (KAIST), a prominent state research institute, said that its researchers developed the self-powered flexible piezoelectric acoustic sensor (f-PAS) for AI voice recognition which can generate power through the tiniest movements and detect different wavelengths.
It is more sensitive than conventional voice recognition sensors, the institute said in a statement.
"The newly developed voice sensor is able to distinguish speakers accurately, so it is possible to apply the technology to a smart home platform or AI assistant," Lee Keon-jae, who led the research, said, adding the sensor can be used as a biometric authentication tool for financial technology (fintech).
Two papers about the sensor were published in the November 2018 issue of Nano Energy, a peer-reviewed academic journal. The institute predicted the global AI voice recognition sensor market would reach $16 billion in 2021.