…continued from Automatic Sound Recognition…

Researchers from Korea University are looking at alternatives to hearing aids and hearing dogs. They hope to employ automatic sound recognition—a technology similar to yet very different from automatic speech recognition, which is key to all interactive voice response, or IVR, systems.

While hearing aids can do wonders, there’s a gap in the coverage they provide. They can’t block out all the background noise, which can confuse hearing-aid wearers (not to mention the rest of us when bombarded with noise).

At the moment, the alternative to hearing aids is hearing dogs—specially trained dogs that can alert their “client” when the phone or doorbell rings, et cetera.

According to DogsfortheDeaf.org, the main purpose of hearing dogs is to give a greater awareness of surroundings for independence and safety. The dogs “are trained to make physical contact and lead their person to the source of a sound.”

The only problem with hearing dogs is that they’re, well, dogs. I’m sure the dogs themselves get a little confused in a crowd. Also, it means the hearing impaired are reliant on someone else (dog, in this case).

What the researchers at Korea University propose is a system that can take the place of a dog and enhance hearing aid use. In-Chul Yoo and Dongsuk Yook published their findings in Automatic Sound Recognition for the Hearing Impaired.

Basically, Yoo and Yook created an algorithm that identifies mechanical sounds such as the phone ringing, the doorbell, a kitchen timer.

Since these mechanical sounds have distinctive spectral peaks, we can classify the spectral bands into relevant and non-relevant groups for each registered sound…non-registered sounds which do not possess such spectral peaks can easily be rejected.

Basically, developers can program the system to identify a library of sounds, reject all others and alert the user when one of the registered sounds occurs. Like a hearing dog—minus the cold, wet nose, I guess (and the affection).