Background noise is one of the main challenges for those with hearing loss because it’s almost inescapable in some form. If you have hearing loss, you’ve probably learned some good tactics for hearing while in the midst of background noise whether you’re at a restaurant with family or shopping with friends. Still, background noise is there and it interferes with hearing.
While hearing aids can sometimes improve background noise depending on the model, they aren’t 100 percent successful, and sometimes sounds are picked up that obscure what the listener is really trying to hear.
But new researchers at The Ohio State University might have a breakthrough. In a recent study published in the journal Acoustical Society of America, Eric Healy, the director of the Speech Psychoacoustics Laboratory at Ohio State, and DeLiang Wang, a professor of engineering and computer science, detail their research using machine learning.
Wang, Healy and graduate student Sarah Yoho tested the hearing of 12 adults with hearing loss by playing speech obscured by background noise. The participants were asked to identify the words they heard while listening without their hearing aids. They then used their specially developed algorithm to remove background noise, and found that when they did this, listening comprehension increased from 25 percent to almost 85 percent.
Then, researchers tested the algorithm on 25 Ohio State students who did not have hearing loss. As it turned out, the scores of those with hearing aids when hearing the algorithm processed sounds were surprisingly higher than those of the students who were not hard of hearing. According to Healy, the implications of these findings are huge:
“That means that hearing-impaired people who had the benefit of this algorithm could hear better than students with no hearing loss,” he said.
The algorithm was developed using machine learning – the science of having computers act without being programmed. In this case, Wang and Healy hope the technology and processors in hearing aids – or eventually, through smartphones – will use artificial intelligence and “learn” to recognize the brain patterns in someone’s deep neural network. The technology would recognize how the brain hears and processes sounds it wants to hear versus background noise and effectively filter these out.
The researchers have partnered with a major hearing aid manufacturer and have received a $1.8 million grant from the National Institutes of Health for further research, which is groundbreaking, according to Wang:
“This is the first time anyone in the entire field has demonstrated a solution,” Wang said. “We believe that this is a breakthrough in the true sense of the word.”
