New ID R&D solution fully automates non-speech monitoring of audio
ID R&D today released Acoustic Event Detection, a new capability that can turn any internet-connected piece of hardware with a microphone, like those in smart speakers, cameras and IoT devices, into a security and safety-monitoring device. The software solution fully automates monitoring of acoustic or non-speech events and, according to the announcement, features an extremely low false positive rate.
Acoustic Event Detection provides continuous monitoring of an audio stream or analyzes an audio snippet and then delivers a result of what type of sound event occurred. The types of events it can detect include a baby crying, glass breaking, gunshots, excessive coughing and sneezing, cries for help, screaming, moaning, and audible alarms. Adding new events is straightforward and can be customized to specific needs.
The AI-based algorithms run embedded on a server or locally on microphone-enabled devices, and it analyzes and returns results nearly instantaneously.
“Integrating our core expertise in artificial intelligence and voice biometrics into a new acoustic monitoring solution is a natural extension for us,” explains Konstantin Simonchik, ID R&D chief scientist. “Just as we provide zero-effort authentication for voice, we now deliver simplicity for new use cases with non-speech sounds. We are essentially offering a virtual human monitor that is 100% attentive, 24/7.”
In January ID R&D opened of an office in Santa Clara, California to be the firm’s new technology development and client support hub.