The software has been fed a database of lung X-rays of thousands of covid, pneumonia and TB patients to tell them apart, the professor says.
New Delhi: An IIT-Roorkee professor claims to have developed software that can detect COVID-19 comparing X-rays of a probable patient’s lungs with thousands of such scans of confirmed cases.
Kamal Jain, a professor at IIT-Roorkee’s civil engineering department, said the software can help in accurate preliminary screening which can be supplemented by clinical testing.
“I first developed an artificial intelligence-based database after analysing over 60,000 X-ray scans, including those of COVID-19, pneumonia and tuberculosis patients to differentiate between the kind of chest congestion suffered in the three diseases. I also analysed the chest x-ray database of the United States’ NIH Clinical Center,” Jain told PTI.
“Using the software I developed, doctors can simply upload pictures of an individual’s X-ray. The software will not only classify whether the patient has any sign of pneumonia, it will be able to tell whether it is due to COVID-19 or (other virus) and also measure severity of the infection. The results can be processed within five seconds” he added.
The professor, who developed the software in about 40 days, has filed a patent and has approached the Indian Council of Medical Research (ICMR) for a review.
So far, there is no verification of his claim by a medical institution.
“The pneumonia caused by COVID-19 is more severe than other (viruses) as it affects lungs completely rather than small portions of the lungs in other cases. The software will analyse bilateral opacity, pattern of fluid build-up in lungs and nature of clump or clot if any,” the professor said.
“Similar kinds of experiments are being conducted by the University of Amazon in the United States but there has not been a breakthrough yet,” he claimed.