Using artificial intelligence algorithms, two engineering students have developed a model that can identify coronavirus infections on X-ray chest radiographs with an accuracy of 97%.The tool is available online Assist healthcare professionals in patient diagnosis.
students Flavio Grillo ÿ Javier Balbas The last phase of the project European University of Madrid A model has been created based on artificial intelligence that can detect COVID-19 infection through chest X-rays.
His research is based on the idea that radiography helps to accurately distinguish pneumonia caused by coronavirus from other pneumonia caused by different pathogens.
“use Convolutional Neural NetworkThe author said that we have applied a very effective learning algorithm to analyze this type of image and can extract various information based on the size and resolution of the input file.
The model has been trained with four types of patient radiographs: Coronavirus pneumonia, a different type of viral or bacterial pneumonia, with healthy lungs, and the fourth type of images failed. After testing, the accuracy rate reached 97.9%.
almost 8,000 X-rays Since they are from the University of Ottawa (Canada), the Italian National Center for Intensive Radiology, a database used by health professionals during the pandemic and shared dozens of x-rays online.
Open to professionals
The tool is freely available to professionals on the web coronavirusxray.com Ø covid19xray.com, Its creator also invites you to share more X-rays to continue to improve the system. Currently, the response has been very positive.
Grillo and Balbás, they also used to University of Hertfordshire (UK) In the scope of the international double degree, they guarantee that their tools are not diagnostic tools, but they do provide support to the medical community, so they can make better diagnoses.
“It can sound an alarm and help health workers who are not necessary Polymerase chain reaction Ø Quick test Coronavirus testing may have doubts about the first clinical impression of patients with pneumonia,” the future engineer concluded.