London: Using artificial intelligence, researchers have created a new smartphone app that may successfully identify COVID-19 infection in people’s voices (AI).
According to them, the technique can be employed in low-income nations where PCR testing are costly and challenging to distribute. The research was revealed on Monday at the international congress of the European Respiratory Society in Barcelona, Spain
.The accuracy of lateral flow testing varies greatly depending on the brand, while the AI model is accurate 89% of the time, according to the researchers.
Furthermore, they claimed that lateral flow tests are noticeably less effective at identifying COVID-19 infection in individuals who exhibit no symptoms. These encouraging findings imply that straightforward voice recordings and refined AI algorithms may be able to identify which patients are infected with COVID-19 with high precision, according to Wafaa Aljbawi, a researcher at Maastricht University in the Netherlands.
These exams are easy to interpret and can be given for free. Additionally, they support remote, virtual testing and have a turnaround time of under a minute, according to Aljwabi.
According to the researchers, the new test may be applied at the entrances to major meetings to enable quick population screening.
The upper respiratory tract and vocal cords are typically impacted by COVID-19 infection, altering a person’s voice. Aljbawi and her managers used information from the crowd-sourced COVID-19 Sounds App from the University of Cambridge, which includes 893 audio samples from 4,352 healthy and unhealthy subjects, of whom 308 had tested positive for COVID-19.
The user’s phone has the app installed. The participants are requested to capture some breathing sounds after providing some demographic, medical, and smoking history basics.
Coughing three times, taking three to five deep breaths through their mouths, and reading a brief sentence on the screen three times are a few of these.
The researchers employed a method for analyzing voice known as Mel-spectrogram analysis, which distinguishes several voice characteristics like loudness, power, and fluctuation across time. Aljbawi explained, “In this method, we can dissect the numerous features of the participants’ voices.
We developed many artificial intelligence models and tested which one performed best at categorizing the COVID-19 instances in order to distinguish the voice of patients with the
disease from those who did not have it, she continued.
Long-Short Term Memory (LSTM) was one of the models that they discovered performed better than the others. LSTM is built on neural networks, which imitate the way the human brain processes and detects the underlying correlations in data.
The researchers discovered that it had an overall accuracy of 89%, a “sensitivity” of 89% for correctly detecting positive cases, and a “specificity” of 83% for correctly identifying negative cases.
In a different study, Henry Glyde, a Ph.D. candidate at the University of Bristol, demonstrated how AI could be used to forecast when patients with chronic obstructive pulmonary disease (COPD) might experience a flare-up of their condition through the use of an app called my COPD. Exacerbations of COPD can be extremely dangerous and are linked to a higher risk of hospitalization.
Breathing difficulty, coughing, and increased mucus production are symptoms.
This article has been directly published from a wire service feed without any text changes. The only change is to the headline.
