Author: David Rodriguez
Coexistence with COVID-19 is now a skill that people have learned. Fortunately, people no longer have to wait in lines to get their COVID tests because of technological advancements. There are now test kits that come in several shapes and sizes.
But wait!
What would happen if a smartphone app could perform a COVID test? Well, that will be so surprising! A study asserts that an application for a phone can successfully identify COVID-19 infection.
Through artificial intelligence, a smartphone app will soon be able to identify persons with Covid-19 infection through their voices precisely. According to PTI, the AI model utilized in the study is more precise than quick antigen testing.
Additionally, they assert that it is inexpensive, quick, and simple to use the new AI model.
The experts claim that this technology can be applied in low-income nations where RT-PCR testing is costly and challenging to perform.
The AI Model is said to be 89% accurate.
According to Wafaa Aljbawi, a researcher at the Institute of Data Science, Maastricht University, The Netherlands, the positive results show that straightforward speech recordings and well-tuned AI algorithms can accurately predict which persons have Covid-19 infection. They have a turnaround time of less than a minute and allow for remote virtual testing.
This method can be employed at the entrances to major meetings to provide quick population screening.
Aljbawi stated at the European Respiratory Society International Congress in Barcelona, Spain.
When working on the research, Aljbawi and her team looked into whether it was feasible to utilize AI to analyze voices to detect COVID-19 infection.
The team is believed to have used data from the crowdsourced COVID-19 Sounds App from the University of Cambridge, which contains 893 audio samples from 4,352 healthy and unhealthy participants.
Three hundred and eight (308) of them had COVID-19 good results. The researcher next used a technique for evaluating voice known as Mel-spectrogram analysis, which determines the number of voice properties like loudness, energy, and variation across time.
We developed various artificial intelligence models and tested which one performed best at categorizing the Covid-19 instances to distinguish the voice of patients with the disease from those who did not, according to Aljbawi. The study’s team found that the Long-Short Term Memory (LSTM) fared better than the others.
The model relies on neural networks, which mimic how the human brain operates and can identify underlying patterns in data.
The team disclosed that 89 percent accuracy was the average for the app. Approximately 89% of positive COVID-19 instances were correctly identified by the app, while 83 % of negative COVID-19 cases were correctly identified.
According to Aljbawi, the findings demonstrate a significant enhancement in the reliability of diagnosing COVID-19 versus current diagnostics such as the lateral flow test.
The New COVID-19 Test Could Be Used at the Entry Point
According to experts, the new Covid-19 test might be applied at the entrances to sizable meetings to speed up population screening. Aljbawi and her team used information from the University of Cambridge’s Covid-19 Sound app, which includes 893 voice samples from more than 4,000 healthy and unhealthy users, of which 308 tested positive.
These tests can be given for free and are easy to interpret. They also allow for remote, digital testing and have a turnaround time of one minute; the scientist was cited as saying by PTI.
Conclusion
Since the pandemic started, scientists have worked hard to find new ways to make COVID-19 testing easy. The use of smartphones in COVID-19 could bring positive changes to simplify the diagnosis.