Researchers around the world are trying to find new and effective ways to detect COVID-19. We¡¯re aware of the swab tests, but the process is pretty uncomfortable while being slightly time-consuming, sometimes taking hours for the results to be officially unveiled.?
However, what if I told you that a researcher in India has found a way to detect COVID-19 within 5 seconds from an X-Ray scan? Sounds good to be true right? Except it is.
Professor Kamal Jain from IIT Roorkee has developed a COVID-19 test that uses the brilliance of AI to detect the novel coronavirus from an x-ray in just three to five seconds.
While being a part of the civil engineering department, Kamal claims to have used data including over 60,000 chest x-rays (sourced from NIH Clinical Centre in the US) to train the AI for detecting the novel coronavirus effectively.
The AI detects the presence of COVID-19 by looking at levels of chest congestion.
He said in a statement, ¡°The usual cause of death in the sample cases I have analysed was severe pneumonia. Most types of pneumonia we know of, for which patients need hospitalisation, are bacterial in origin, and respond to antibiotics. Pneumonia caused by COVID-19 is particularly severe, and tends to affect all of the lungs, instead of small parts.¡±
He added, ¡°Identifying bilateral opacities, the pattern of fluid build-up in lungs, nature of clumps and its overall arrangement is the main source of identification, but it takes time and experience for physicians who know how to look at the data. But with our AI-based app, we can quantitatively classify such patterns in no time, and that too for thousands of patients.¡±
Professor also hinted that the AI can not just detect the presence of COVID-19, but also the severity of the case, thus indicating to the doctors in case he/she might need special attention, stating, ¡°The main advantage of this app is that it is able to process lakhs of x-ray images in a matter of minutes, and classify a patient according to the level of coronal infection, to help hospital administration decide which patient needs to be put on ventilator first.¡±
He added, ¡°Thus, with this app, we can not only identify which patient is corona-infected, but also determine the severity of infection with visualisation of those parts of the lungs that are infected, and add a level of smart management by automatically arranging patients who need to be ventilated in a queue.¡±?
While the AI system seems like a novel way that could help many, it is yet to get approved by ICMR (Indian Council of Medical Research) and Jain is anticipating its review. It hasn¡¯t been approved by any medical institution across the world as of now.