{"id":1013,"date":"2022-04-21T21:10:24","date_gmt":"2022-04-22T01:10:24","guid":{"rendered":"https:\/\/silverspringhearing.com\/?p=1013"},"modified":"2022-04-21T21:10:29","modified_gmt":"2022-04-22T01:10:29","slug":"machine-learning-enhances-speech-recognition","status":"publish","type":"post","link":"https:\/\/silverspringhearing.com\/machine-learning-enhances-speech-recognition\/post","title":{"rendered":"Machine Learning Enhances Speech Recognition"},"content":{"rendered":"\n
As researchers work to understand more about the impact of hearing loss<\/a>, one area of particular interest is the way it affects our ability to recognize speech. A new study created a human speech recognition model based on deep machine learning that produced some promising results.<\/p>\n\n\n\n In sensorineural hearing loss (the kind caused by age or exposure to loud noise) people tend to struggle to hear high-frequency sounds first. This includes voices that are higher-pitched and most consonant sounds. This makes understanding speech much more difficult. Certain circumstances can make it even harder, including:<\/p>\n\n\n\n Because of this, people with hearing loss struggle in social situations like large gatherings or dinner with friends at a busy restaurant like Copper Canyon Grill.<\/a> This can lead to frustration, anxiety and even isolation for those struggling with their hearing.<\/p>\n\n\n\n Hearing aid algorithms are used to help improve speech recognition in people with hearing loss. However, current evaluation methods for these algorithms are costly and time-consuming.<\/p>\n\n\n\n To help combat that, researchers in Germany conducted a study using automatic speech recognition (ASR) to calculate how many words per sentence listeners understood.<\/p>\n\n\n\n The study consisted of eight participants with normal hearing and 20 participants with hearing loss who were exposed to a variety of complex noises that mask speech<\/a>. Researchers separated those with hearing loss into three different groups based on the severity of their condition.<\/p>\n\n\n\n They found that their model was successful at predicting human speech recognition performance across all noise types and for different hearing abilities.<\/p>\n\n\n\n The hope is this model can be used to improve hearing aid algorithms and help users with speech recognition in any environment.<\/p>\n\n\n\nHow Hearing Loss Affects Speech Recognition <\/h2>\n\n\n\n
<\/figure><\/div>\n\n\n\n
Research Seeks to Improve Hearing Aid Algorithms<\/h2>\n\n\n\n
Don\u2019t Leave Your Hearing Loss Untreated<\/h2>\n\n\n\n