IIT Mandi-Led Team Develops AI App To Track Home Quarantined COVID Patients
A team led by researchers at the Indian Institute of Technology (IIT) Mandi has developed an artificial intelligence (AI) based biometric application to continuously monitor and accurately detect the identity of a home quarantined COVID-19 patient.
A team led by researchers at the Indian Institute of Technology (IIT) Mandi has developed an artificial intelligence (AI) based biometric application to continuously monitor and accurately detect the identity of a home quarantined COVID-19 patient. The mobile application, dubbed LakshmanRekha, uses a combination of biometric verification, geofencing and AI so that no patient can breach the assigned quarantine space. Geofencing technology uses GPS signals to determine when a person enters a certain boundary.
"We have developed a pilot version of LakshmanRekha mobile application and tested it over small datasets," Aditya Nigam, Associate Professor, School of Computing and Electrical Engineering, IIT Mandi, said.
"The obtained results are very good and now we are working to add more functionality, scalability and usability to make it ready for deployment,” Mr Nigam said.
The results of the research, funded by the Department of Science and Technology, have been published in the IEEE Consumer Electronics Magazine.
The team noted that in existing quarantine management mobile applications, individuals under self-isolation are enforced to share their instantaneous position routinely via geofencing technology. They are also required to upload a face selfie every hour or ten times a day, according the, researchers, including those from IIT Delhi and BITS Pilani, Rajasthan. However, these geofencing applications fail to ensure the user identity throughout the time because individuals can leave cell phones in isolation zones and move in or out, leading to breach of the self-isolation rules, the researchers noted.
The idea of uploading a face selfie every hour also cannot ensure the patient’s stay in a geofenced area as they can also try to fool the system by using a photo containing its registered face, they said. To counter these limitations, the new app developed by the team matches the quarantine location of the individual with the place from where they have uploaded the biometric data.
Using AI, the application continually computes an authentication score that can measure how certain it is that the quarantined user is also the one using the mobile, the researchers said. If the application detects any action indicating that the user identity has been changed, it will directly notify the authorities for necessary action, they said.