A new smartphone app created by researchers may successfully identify COVID-19 infection in people’s voices using artificial intelligence (AI). Based on the researchers’ view, the AI model employed in the study is less expensive, quicker, and easier to use while being more accurate than rapid antigen testing or lateral flow tests. Researchers say that a new smartphone app uses artificial intelligence to properly detect Covid-19 in people’s voices.
Aim of building the AI-based App
The research was revealed on Monday at the international congress of the European Respiratory Society in Barcelona, Spain. So according to them, the technique can be employed in low-income nations where PCR testing is costly and challenging to distribute.
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.
Features of the App
As per Wafaa Aljbawi, a researcher at Maastricht University in the Netherlands, the features implemented in this application are
- Straightforward speech recordings and fine-tuned AI algorithms can reach high precision in diagnosing which individuals have COVID-19 infection.
- These tests are easy to understand and can be given for free.
- They also support remote, virtual testing, having a one-minute turnaround time.
- The new test may be applied, for instance, at the entrances to major meetings to enable quick population screening, according to the researchers.
- The upper respiratory tract and vocal cords are typically impacted by COVID-19 infection, altering a person’s voice.
Aljbawi and her associates 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.
What are the steps for these AI-enabled COVID-19 tests?
- 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.
-
Mel-spectrogram analysis
The researchers employed a method for analyzing voice known as Mel-spectrogram analysis, which distinguishes several voice characteristics like loudness, power, and fluctuation across 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 to distinguish the voice of patients with the disease from those who did not have it.
A spectrogram that has been scaled to the Mel scale is called a Mel spectrogram. The frequency spectrum of a signal is the frequency range that the signal contains, and a spectrogram is a visual representation of the frequency spectrum of a signal.
What is the outcome of the recent study?
The encouraging results show that Long-Short Term Memory (LSTM) was one of the models that they discovered that performed better than the others. Neural networks, which replicate the way the human brain functions and recognize the underlying correlations in data, are the foundation of LSTM.
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.
Previous Study on this AI Technology
Earlier in the year 2020, students in Mumbai create a voice-based AI tool to identify COVID-19. Three biotechnology students and a professor from Mumbai developed a patented artificial intelligence-based tool, which they say can test COVID-19 through voice-based diagnosis using a smartphone.
The tool is being evaluated by the University of Tor Vergata in Rome, and 300 people have already used it.
The technology is based on a voice-based diagnostic via an app, said the students and professor from Mumbai’s DY Patil Institute of Bio-Technology and Bio Informatics.
Bioinformatics students Rashmi Chakraborty, Priyanka Chauhan, and Priya Garg are on the team.
About the Research
The students’ software is fully functional and has a large database of patient and healthy sample data. With 98% accuracy, the University of Rome is currently using this technology to identify COVID-19 patients, according to professor Santosh Bothe, who oversaw the project.
The program analyzes voice metrics including frequency and noise distortion as a user talks into the app’s microphone. The proprietary technique then determines whether the patient is positive or negative by comparing these numbers to those of a normal person.
Working on a diagnostic tool based on the analysis of cough and respiratory sounds is a team from the Indian Institute of Science (IISc), Bangalore.
The audio-based disease detection tool may detect coronavirus from the timbre of the voice, claims Giovanni Saggio, professor at the Rome university’s Engineering department.
“Each human voice has 6,300 parameters, yet fewer than a dozen of them specifically identify individuals. The human ear cannot tell the difference, except for colds, but artificial intelligence can. Every one of our internal organs acts as a resonator, so if we have an issue with our heart or lungs, it will be audible in our voice, according to Saggio.
“The same person speaks with one voice when they are healthy and a different voice when they are ill. Because the coronavirus damages the lungs and airways, the voice is undoubtedly impacted. This method might be used to select the current novel coronavirus cases, he continued.
According to Priya Garg, a student researcher, this tool can have a significant impact on the initial screening process for identifying positives. Only those who tested positive can proceed with the lab testing. At that time the aim was to assist the medical infrastructure, to assist the government, in identifying hotspot zones beforehand through location tracking integration, and is perhaps the finest approach to reach out to the most rural portion of India by testing through a smartphone.
Aarogya Setu App
This app’s declared goals are to raise awareness of COVID-19 and to link Indians to crucial COVID-19-related health services. By disseminating best practices and cautions, this app supports the Department of Health’s initiatives to manage COVID-19.
The app acts as a tracking application that monitors COVID-19 cases by utilizing the GPS and Bluetooth capabilities of the smartphone. Both iOS and Android mobile operating systems support the app. By scanning a database of reported cases across India, Bluetooth tries to assess the risk if a user has been close to (within six feet of) a COVID-19-infected individual. Based on the information provided, it assesses whether the location one is in is a part of the infected areas using location information.
This Indian tool is completely functioning and is currently being used in Italy to identify COVID-19 patients, while other foreign universities are attempting to establish a voice-based AI tool for COVID-19 identification.
Conclusion
With the invention of modern technology and its incorporation into medical research, we can now expect more progress in assessing our health and hygiene in a smarter and faster way.
.){kind=link}