Computer Vision can help spot Covid-19 in chest X-ray
The coronavirus disease (COVID-19) continues to affect communities as it spreads and threatens the health of the people across the world. For months now, the healthcare sector suffers from the surge of patients needing care.
Many technologists grapple and think, “How can we help?”
Computer vision for early detection
In diagnosing the disease, infected patients are screened to be given proper treatments, and one of the screening methods is analysing the chest radiograph or x-ray, which can be too much for doctors and nurses.
Artificial Intelligence firm DarwinAI and experts from the University of Waterloo in Canada thought they could help.
They developed a neural network called COVID-Net, which utilises computer vision to scan images for signs of the virus. It was trained with 5,941 chest x-ray images from 2,839 patients with lung viral and bacterial infections, including COVID-19.
A call for a community-driven AI
The founders of the neural network open sourced their model to the community “in hopes of developing a robust tool to assist healthcare professionals in combating the pandemic.”
The COVID-Net source code, data sets, documentation, and other pertinent information needed to develop a solution can be found in a GitHub repository.
Just like other successful computer vision solutions, rigorous training and development is needed to complete a robust tool for early detection of COVID-19.
The journey continues...
A lot of work has to be done, but the COVID-Net initiative is a great start. Researchers and AI professionals can explore the neural network to make it work.