Using person’s sound to detect infectionDate: 10 April 2020 Tags: Fourth Industrial Revolution
Researchers from the University of Cambridge have designed a new mobile phone app that will be used to collect data to develop machine learning algorithms, which will be able to detect if the person has COVID-19 or not based on the sound of their voice, their breathing and coughing.
Since COVID-19 is a respiratory condition, the sounds made by people, including those made during breathing and coughing will be very specific, creating a possibility for diagnosis externally.
Users of the app have to take a survey, which will collect their basic demographic information and medical history and takes their sound recordings, along with samples of their breathing and coughing. The aim of the research is to collect data from both healthy and non-healthy people.
Once the researchers have collected a large database of such information, they will release it to other researchers and they hope that the dataset will help shed light on disease progression and the relationship between the respiratory complication arising out of COVID-19 with the medical history. The app will also ask if the person has tested positive for COVID-19.
Artificial Intelligence (AI) is when machines are trained to perform human tasks. Machine learning, on the other hand, is a subset of AI and involves training machines to look for patterns in datasets, something which humans would take much longer to identify.
To do this, however, machines need to be provided a considerable amount of data that they can comb through to figure out any patterns in it, which is the reason why researchers from Cambridge aim to collect a large enough data set of sounds for the machines to work on it.
Some of the other AI-based approaches to deal with the current pandemic include the COVID-19 Open Research Dataset (CORD-19), which is a free resource of the research that is being done on COVID-19.
There is also the open-access neural network, COVID-Net, which aims to help researchers around the world develop an AI tool that can test people for COVID-19. The COVID-Net is being trained by researchers to identify signs of coronavirus in chest x-rays using roughly 6,000 images taken from over 2,800 patients with various lung conditions.