COVID-Net can help doctors detect Coronavirus in chest x-rays

Original article can be found here (source): Artificial Intelligence on Medium

Example chest radiography images of COVID-19 cases from 2 different patients and their associated critical factors (highlighted in red) — Credit: COVID-Net Study

COVID-Net can help doctors detect Coronavirus in chest x-rays

This convolutional neural network (CNN) can help researchers develop an AI tool to efficiently pick up signs of the disease

Although most of the Artificial Intelligence systems are still in the stage of infancy, they have proved their usefulness in a variety of trials across the wide spectrum of the ecosystem. And a lot of people are wondering why we are not utilizing the technology to help tackle the current outbreak. The simple answer is that it could lead to ill-informed decisions where public money is spent on unproven AI technology.

Having said that AI has shown great promise in the healthcare sector with the prediction, diagnosis, and treatment of various diseases. An interesting perspective with regards to the current pandemic is that an AI company called BlueDot, which uses machine learning to monitor outbreaks of infectious diseases, was the first to alert its clients — governments, hospitals, and businesses, of an outbreak in China on December 30th.

The company which monitors outbreaks of infectious diseases around the world reported a spike in the pneumonia cases in Wuhan, China — the epicenter of COVID-19. This was a full nine days ahead before the World Health Organization (WHO) officially warned about the disease.

Two more similar services, namely HealthMap at Boston Children’s Hospital and Metabiota, based in San Francisco also caught those early signs. Although human teams say they spotted the disease the same day as these AIs, it is still impressive how these automated systems picked up an outbreak across the world.

In a piece of encouraging news, an open-access neural network dubbed as COVID-Net was recently released to the public. The endeavor is meant to help researchers around the globe in developing an AI tool that can be used for effectively testing people suffering from the COVID-19. For those of you who are new to this, a convolutional neural network is a type of AI which is really good at recognizing images.

The neural network was designed by Linda Wang and Alexander Wong at the University of Waterloo in Canada. Collaborating with a local AI firm DarwinAI, the tool was trained to identify the signs of COVID-19 by analyzing the chest x-rays of the patients. As you may already know, the current pandemic is primarily a lung infection.

COVID-Net looked at 5,941 images taken from 2,839 patients with various lung conditions, including bacterial infections, non-Covid viral infections, and Covid-19 to differentiate the pandemic from other diseases. This data set has also been released publicly along with the tool to prove its credibility and for the researchers to tweak it whichever way they want to.

However, this is not the first CNN developed in this area. Several other companies have announced their own AI tools which can perform the same function. But the major difference between them and COVID-Net is that only the latter has made its tool and findings completely public. CEO Sheldon Fernandez of DarwinAI was quick to point out that it is not a “production-ready solution”, but would need others’ help to convert it into one.

While AI may not be able to save us from the coronavirus, but the promise of such futuristic tools would certainly play a much bigger role in any future outbreaks.

Complete research was published in the Preprint archive Arxiv.