Millions of hand-drawn hearing tests from before 2005 are locked away on paper, unable to be analysed by modern computers. This matters because audiograms—graphs showing how well a person hears different pitches—are the standard record of hearing health. Without a way to digitise the older paper versions, researchers cannot include decades of patient data in large-scale studies. That gap makes it harder to track how hearing changes across a lifetime, predict who will need early intervention, or develop better treatments. The problem is not just historical: many community clinics still record tests by hand. The researchers will build an artificial intelligence system that can automatically "read" hand-drawn audiograms and convert them into digital data. They will test it on over 600,000 hearing tests to ensure accuracy, then release the tool online for free. If it works, hospitals and researchers worldwide will be able to digitise their archives, giving studies access to far more representative data—covering all ages and backgrounds, not just those who had digital tests after 2005. This could help identify patterns of hearing loss progression and support earlier, more targeted interventions.
View original technical description
Background When you have a hearing test, the results are drawn as a graph called an audiogram, showing how well you can hear sounds at different frequencies (or pitches). Since 2005, these tests have been recorded digitally on computers. However, millions of hearing tests taken before 2005 were drawn by hand on paper. These older tests, along with many current tests done in community clinics, contain valuable information about how hearing changes over time – but because they're on paper, they can't be analysed using modern research methods. This matters because understanding how hearing changes throughout someone's life is crucial for developing better treatments and predicting who might need early intervention. For example, comparing how someone's hearing looked at age 30 with how it appears at age 70 could help doctors understand how hearing loss progresses and identify the best ways to protect hearing. Aims In this project, the researchers will develop a new computer system that can automatically convert these hand-drawn hearing tests into digital information. They'll use artificial intelligence to teach computers how to "read" the graphs and convert them into numbers that researchers can analyse. They’ll test their new system on over 600,000 hearing tests to ensure it works accurately and reliably. Once they’ve developed it, the tool will be made freely available online so that hospitals and researchers worldwide can digitise their historical hearing tests. This means that for the first time, researchers will be able to include older records in modern research, making their understanding of hearing loss more complete and representative of the whole population. Benefit This research could benefit people with hearing loss in several ways: It will help doctors and researchers better understand how hearing changes over time, leading to more accurate predictions about hearing loss progression. It will ensure that research includes data from people of all ages and backgrounds, not just those who had digital tests. It will give researchers access to much more information about hearing loss, helping them develop and test new treatments. It will help identify patterns that could lead to earlier intervention and better protection of hearing.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know