A study by researchers at the University of Manchester shows that AI-powered hearing tests allow you to check your hearing on a computer or smartphone without clinical supervision.
By using AI to screen for background noise, they say high-tech hearing tests allow people to efficiently understand human speech from the comfort of their own home rather than from a hospital clinic.
Researchers have developed and tested an AI-powered version of the DIGITS-IN-NOISE (DIN) test, combining Text-to-Cheese (TTS) with automated speech recognition (ASR) technology.
The result was a fully automated, self-controlled hearing test that could be performed in 10 minutes without clinical supervision.
Funded by the Medical Research Council's PhD training partnership grant, the study could revolutionize the way hearing testing is performed, and is published in today's journal Trends in Hearns.
“We've tested this technology, and we're confident that with the help of AI, we can automate hearing tests on a computer or smartphone, from the comfort of our home,” said Mohsen Fatehfer, lead author of the University of Manchester.
“While we still need a wider range of testing and a user-friendly interface, this technology can make a huge difference for patients.
“The specialist equipment in the clinic and the specially trained staff needed to use it are not always available to patients who need a rapid assessment.
“And more, when experiencing hearing impairment, people are slow to ask for help. There is an estimated delay of 8.9 years while hearing aids are needed during the time of adoption.
“That's why we're excited by the ability of this system to incorporate machine learning into our testing procedures and reduce our dependence on human supervisors.”
Speech-in-noise tests are commonly used to detect hearing impairment problems by assessing how well you can understand the spoken speech rather than background noise.
Traditional tests usually rely on pre-recorded human speech, and require clinicians to score responses.
However, the AI-powered version will be replaced by both computer-generated voice recognition and automated voice recognition to ensure that the tests are fully performed.
In a group of 31 adults, some with normal hearing and hearing loss, AI-powered tests were evaluated against two traditional DIN tests.
The researchers evaluated the reliability of both the degree of consistent outcomes across multiple runs and validity.
The results showed that AI-powered tests yielded roughly the same results as traditional DIN tests.
In some cases, there was slight variability, especially for people with strong accents, but overall reliability and accuracy were the same, indicating that the addition of AI had no negative impact on test performance.
Also, by using a larger ASR system, researchers say that higher accuracy makes the system compatible with stronger accents.
Co-authors, Professor Kevin Munroe and Michael Stone, graduates from the University of Manchester and are supported by the National Institutes of Health (NIHR) Manchester Biomedical Research Center.
Professor Munro said: “This study highlights how AI performs hearing tests both reliable and user-friendly, especially for individuals who may use traditional formats such as keyboard and touchscreen challenges.
“It also illustrates an important step towards a more personalized, accessible hearing assessment that allows people to complete independently at home.
“Test software is freely available and provides the foundation for future development using more advanced voice technology.”
Professor Stone said:
“Our team will explore expanding this technology into more complex speech testing in future research.”
/Public release. This material of the Organization of Origin/Author is a point-in-time nature and may be edited for clarity, style and length. Mirage.news does not take any institutional position or aspect, and all views, positions and conclusions expressed here are the views of the authors alone.
