Two third-year computer engineering students from Pune Institute of Computer Technology (PICT) were the runners up in the global category of the International Better Health Hackathon.
Krishna Bhambhani (19) and Tanmay Jain (20) participated in the category for their solution CoVidSpy at the Better Health Hackathon: #CodeForCovid19, and bagged the runners up position, in addition to winning USD 10,000. The international event, which concluded on August 28, was organised by HC Technologies, in collaboration with Microsoft. The team’s solution stood second best among solutions from more than 200 teams and 7,500 participants across over 50 countries.
“We started our project mid-May, when Covid-19 cases were on the rise and several countries were coming out of lockdown,” Bhambhani said. “Organisations took measures to ensure (enforcing) rules like physical distancing and wearing masks to reduce the risk of infection, but this was not executed well in real time. There were also systems designed to aid detection, but this could not happen efficiently on local computers or machines.”
She said CoVidSpy is a pandemic management system that is artificial intelligence (AI) and machine learning enabled, along with being computer vision based. The primary objective is the enforcement of safety norms in public organisations and institutions.
The product is named CoVidSpy – ‘Co’ encourages cooperation among people in the fight against the pandemic and ‘Vid’ indicates that it is primarily a video-based management system. “We named it CoVidSpy as we will essentially be “spying” on the virus to ensure members of the community are not infected,” Jain said.
CoVidSpy has four main modules – fever detection, mask detection, physical distancing violation detection, and dividing a room into zones to detect areas where distancing violations are happening, which is referred to as zoning. With the help of an object detection algorithm, CoVidSpy will help organisations identify faces with and without masks, and has computer vision-based temperature estimation using infrared camera to determine whether or not a person has fever.
“…If an individual is detected violating physical distancing with respect to another individual, a zone alert will be triggered, which will be a visual indication to alert people that they are not practising the recommended precautions…” Bhambhani said.
The duo said mask detection would not reveal any personal information of any individual. “CoVidSpy does not use facial recognition, and does not store any data of individuals. It simply detects violation to aid authorities at localised levels,” Jain said.
Under the guidance of Dr Kavita A Sultanpure, as well as the management and director of PCIT, Bhambhani and Jain are optimistic about their frontend and backend system, which has the potential for usage across countries affected by the pandemic. The duo was also given Rs 1 lakh from PCIT.
The duo is currently working towards magnifying the system. “We evaluated places that will be frequented by people as soon as the lockdown is revoked,” Jain said. “For now, our target is retail spaces, and we are also looking into banks and government offices…We are working towards making the system more extensive, yet easy to operate, with multiple cameras and high computational power, so large public spaces, including roads, can use it.”
📣 The Indian Express is now on Telegram. Click here to join our channel (@indianexpress) and stay updated with the latest headlines