IIT Jodhpur develops AI method that highlights COVID affected lung area using X-ray images

A team of researchers from IIT Jodhpur have developed an automated Artificial Intelligence (AI) solution that can differentiate a COVID-19 infected lung from a non-infected one using X-ray images. According to a report by news agency ANI, the solution is based on a deep-learning-based algorithm called COMiT-Net which studies chest X-ray images and differentiates between a healthy and an infected lung by pointing out the abnormalities. It is worth mentioning that detecting COVID-19 using X-ray images has been an in-use technique however this technology promises a unique advantage. 

AI algorithm delivered results with 96.8% sensitivity

The new technology was tested in an experiment wherein the experts analysed 2,500 chest X-ray images and achieved 96.8% of sensitivity. While there are other techniques to diagnose the presence of the Coronavirus using X-ray images, IIT Jodhpur developers said that their method gives an extra edge over others. They claimed that their AI technology is able to accurately locate the infected region in the lungs which makes it 'explainable'. 

"The developed algorithm not only predicts whether the CXR has COVID-19 pneumonia or not, but it is also capable of identifying the infected regions in the lungs, thus making them explainable", IIT Jodhpur said in its statement as per ANI. It must be noted that there have been several kinds of research to detect COVID-19 using X-rays and CT scans, no method has been able to provide a visual picture of the infected area. Underlining that their technique interprets only from the lung region, IIT Jodhpur said that their AI solution is explainable for both points of view- algorithmic and medical. 

The developers have published their research, which is a part of the RAKSHAK project under NM-CPS DST and iHuB Drishti at IIT Jodhpur, in the journal Pattern Recognition. In the steps ahead, the experts are now preparing to further develop this technology by doing follow-up studies on the results of their AI project. 

 Image: ANI