With the arrival of winter in Europe, concern for Lung diseases like pneumonia keeps hospitals on alert. A group of researchers proposes the use of Artificial intelligence to help detect this type of evil: the system would work with 98% accuracy.
Scientists from the University of West Scotland It would help to diagnose with high efficiency, almost automatically, lung diseases.
Currently, his diagnosis It would require various tests, including X-rays, blood tests, ultrasounds, and CT scans.
But these tests take a long time before seeing the definitive results, in addition to being expensive.
The University of West Scotland created Artificial Intelligence software, initially, for the detection of COVID-19. Now it would be adapted to other lung diseases, such as the already named pneumonia, in addition to tuberculosis.
Naeem Ramzan, a professor and researcher at UWS, is at the forefront of the investigations.
“Systems like this could prove crucial for medical teams around the world. There is no doubt that hospital departments around the world are under pressure, and the COVID-19 outbreak has exacerbated this, adding further strain to departments and staff,” said Ramzan.
“There is a real need for technology that can help alleviate some of these pressures, detecting a range of different diseases quickly and accurately, which helps free up valuable staff time.”
This is how Artificial Intelligence works to detect lung diseases
The team created by the UWS uses x-ray images, comparing it with a database of thousands of images of patients with pneumonia, tuberculosis and COVID-19.
Let us remember that Artificial Intelligence is the attempt to recreate human intelligence in machines or computers. It’s done through training, with machine and deep learning.
In this case, the training consists of the software analyzing thousands of X-ray images of patients.
Using the deep convolutional neural network, images are compared, determining the existence or not of any disease.
During the study phase, as highlighted by the portal Interesting Engineering, the technique worked with an accuracy of 98%. Now, researchers at the University of West Scotland hope to spread the method to other hospitals, to better determine its accuracy.