AI-powered Mobile App “Trained” to Analyse Cough and Make Diagnosis
It was only late last year when it was announced as a trial app. Australian scientists have developed the smartphone app called ResApp that can analyse the sound of a child’s cough and diagnose whether it is asthma, croup, pneumonia and other respiratory illnesses.
Today, Aussie researchers at Joondalup Health Campus and Princess Margaret Hospitals who developed the app are already using it to detect respiratory diseases in children.
The AI-powered smartphone app can perceive and analyse respiratory illness by “listening” to a person’s cough. The app has an algorithm that bases its analysis on the features of the cough that are distinct to five different diseases. Typically, this procedure is performed by a doctor when he listens to the sound of breathing or coughing using a stethoscope.
To create this app, the researchers had to collect a database of cough audio recordings from 1473 hospitalized children who were 29 days to 12 years old and confined to a hospital. Using iPhone 6 smartphones, the scientists recorded the cough audio streams in realistic hospital environments, meaning in the presence of background noises such as talking, crying, footsteps, equipment running, doors closing, and the like.
Then, through machine learning algorithms similar to speech recognition systems, the app was trained to identify the specific cough sounds associated with pneumonia, croup, asthma, bronchiolitis, and general lower respiratory tract disease with 852 of the recordings. Next, the app was tested on the remaining 585 children and it was found to have an accuracy rate of 81 to 97%.
According to Associate Professor Udantha Abeyratne, a University of Queensland biomedical engineer who helped develop the app, “Coughs can be described as wet or dry, brassy or raspy, ringing or barking; they can whistle, whoop or wheeze; but experts cannot always agree on the description or how to use cough sounds for diagnosis.” He added that the app is a big help in early diagnosis and better patient outcomes, especially in remote areas where access to medical facilities are limited.
Co-author Dr. Paul Porter also proposed that the technology is beneficial in terms of identifying the common respiratory disorders without having to go to the doctor for clinical examination, X-rays or bronchodilator testing. He also adds that the research team is already expanding the study to look at adult respiratory diseases and explore the technology of machine learning in order to pin point the level of disease severity.
The development of the app means a medical diagnosis can be made using only a smartphone, acoustic analysis and the power of mathematics. The researchers published a paper on the app in the open access medical journal Respiratory Research.
Jaclyn-Mae Floro, BCompSc
Contact W3IP Law on 1300 776 614 or 0451 951 528 for more information about any of our services or get in touch at firstname.lastname@example.org.
Disclaimer. The material in this post represents general information only and should not be taken to be legal advice.