Machine-learning technology that mimics the way the human ear works and "listens" for potential dangers in the surrounding environment is being developed by the startup One Llama.
According to MIT Technology Review, the app, called Audio Aware, is set to launch in March and is aimed at smartphone users with hearing problems or severe…absence of sound. It will run in the background of Android and will have an “ear” for sounds like tires, sirens, etc., informing the user when it detects something (for example, by pausing the music).
The app will feature a database containing a wide range of "dangerous" sounds, and users will be able to add their own.
One Llama is targeting the emerging wearables market, with the technology set to be incorporated into smart glasses, watches and fitness devices. Audio Aware could do more than just alert people to dangers, such as monitoring health status or even “detecting” the wearer’s location by “listening” to surrounding sounds.
At the heart of this technology is what the company calls an “artificial ear” – essentially software that emulates the way the human ear works, based on the work of David Cheng, the company’s co-founder, and others at the University of Illinois.
As Cheng explains, Audio Aware works by “listening” through the smartphone’s microphone, comparing the sounds in the environment with those it has stored in its database. When it detects a match – such as, for example, a car horn – it then stops any sound the user is currently hearing from their phone and plays it back, increasing its volume or making it more distinctive and easier to hear.
Audio Aware will be able to work without access to wireless networks, but an Internet connection will be required for "learning" and updating for new sounds (such as if the user goes to another country, where the sirens are different).
Source: naftemporiki.gr
