Health
COVID-19 Can Be Detected Via Mobile Phone App Through a New AI Model
(CTN News) _ An AI (artificial intelligence) model could detect COVID-19 through someone’s voice through a mobile phone app, according to a recent study presented on Monday at the European Respiratory Society International Congress in Barcelona.
This AI model is cheaper, faster, and easier to use than lateral flow/rapid antigen testing, so it can be applied in low-income nations where PCR tests are costly or difficult to obtain.
89% Accurate
According to Wafaa Aljbawi, a researcher at Maastricht University in the Netherlands, lateral flow testing varies greatly depending on the brand. The AI model was accurate 89% of the time.
In those without symptoms, lateral flow tests had a significantly lower ability to detect COVID-19 infection.
Based on these encouraging results, basic voice recordings and customized AI algorithms may be able to accurately identify patients infected with Covid 19.
Aljbawi claims such tests are easy to interpret and can be delivered at no charge.
Voice quality is affected by Covid 19 in the upper respiratory tract and vocal cords.
Visara Urovi and Sami Simons from Maastricht University Medical Centre tested whether voices could be used to identify COVID-19.
The University of Cambridge provided 893 audio samples from 4,352 healthy and unhealthy subjects, of which 308 tested positive for COVID-19.
How the App Works
In order to use the app, users must download it to their smartphones, provide some basic demographic, medical, and smoking status information, and then record respiratory sounds.
To get tested, users must cough three times, take three to five deep breaths, and read a brief sentence three times.
A method for analyzing voice called Mel-spectrogram analysis was used, which distinguishes several voice characteristics like volume, power, and fluctuation.
One of the models they discovered that performed better than the others was Long-Short Term Memory (LSTM).
LSTMs imitate how human brains function and identify the underlying links in data through neural networks.
According to News Medical, because it can store information and works with sequences, it is suitable for simulating signals gathered over time, such as speech sounds.
Further Validation of Findings
In order to validate their findings, the researchers said substantial numbers should be used.
To enhance and validate the model’s accuracy, 53,449 audio samples have been collected from 36,116 individuals since the experiment began.
The researchers COVID-19 are also conducting additional research to determine which aspects of the speech are influencing the AI model.
SEE Also:
Labor Day Hours At Costco In 2022: What You Need To Know
Cyber Police Arrest 6 Facebook Loan Fraudsters in Chiang Rai
