Local software can ‘detect Covid within a minute’

Algorithm that uses chest X-Ray to diagnose disease gets DRAP approval

ISLAMABAD:

The Drug Regulatory Authority of Pakistan (Drap) has granted approval to "Cov-raid", a locally developed software capable of “detecting the novel coronavirus in a person's lungs within a minute".

Kudos to NECOP and Pakistani scientists for developing ‘COVRAID’ an AI based diagnostic software to detect COVID-19. Pakistan stands ready and willing to explore collaborations and partnerships to share this technology with the world (1/2)#Innovation #AI4Health #PakistanScience pic.twitter.com/t0X1MByPv8

— Science Diplomacy Pakistan (@SciDipPak) November 20, 2020

 

Cov-raid (Rapid Artificial Intelligence Detection), developed by the National Electronics Complex of Pakistan, requires a chest X-Ray image as an input for the detection of Covid-19.

“The local invention of Cov-raid is a major achievement,” DRAP CEO Dr Asim Rauf said on Saturday. “It will have a significant contribution,” he added.

The official said the technology was available in only a few countries in the world. “Pakistan will supply the Cov-raid to other countries too. It will soon be available in the entire country.”

The Cov-Raid has been registered under the DRAP Act 2012. It will “employ Convolutional Neural Networks” to diagnose Covid-19 in suspected individuals by using X-rays and it has been approved for "secondary detection" of the virus.

According to the Cov-Raid website, the artificial intelligence software has been developed by creating a data repository of chest X-rays for Covid-19 detection.

It highlights its finding in lungs through an AI approach by using heatmaps. The algorithm has been trained on more than 35,000 chest X-rays with data authentication carried out through multiple certified radiologists and PCR reports.

“Our system analyses chest X-ray images in seconds and [uses] suspicious lesions and abnormalities to help diagnose pneumonia caused by Covid-19. The model, in particular, is optimised in detecting interstitial opacity, effusion, and nodule based on data of confirmed patients,” the text on the website read.

“The existence of portable chest X-ray machines [in remote areas/towns/cities/isolated places will enable] reducing the risk of Covid-19 transmission during the mobility of a patient,” it added.

“Furthermore, portable setups for chest X-ray can be quite effective in determining the geographic clusters of COVID-19 and patients may be instructed to stay at home for effective disease control, until the onset of advanced symptoms.”

Last month, researchers at the University of Minnesota, working with Epic Systems Corporation, said they had validated an artificial intelligence algorithm that could assess chest X-rays for potential cases of Covid-19.

To develop the algorithm, a team of University of Minnesota experts led by Assistant Professor Ju Sun, analysed de-identified chest X-rays.

To train the AI to diagnose Covid-19 specifically, the researchers used 100,000 X-rays of patients who did not have the virus and 18,000 X-rays of patients who did.