AI Based Coronavirus Detection for Corona Virus Screening

How Artificial Intelligence can help fight Corona Virus

How we use thermal cameras and Artificial Intelligence to detect people who have higher body temperatures, and hence screen them out to prevent further spread of the virus

With the number of coronavirus cases in India rising at a rapid pace, there exists a dire need for more testing to determine the exact extent of the problem. It’s no secret that there is a shortage of testing kits in India and while that shortage exists, there needs to be other ways to check if someone is affected. One of the ways of doing that is checking if a person has a higher than normal body temperature which means that he/she has fever and that could be an indicator of that person being affected.

Typical ways of doing that is using a thermometer either contact-based or contactless. The drawback is that the detection speed is low, people need to get tested one by one and there is a significant chance of contact even while using a contactless thermometer. 

To combat that problem we have developed a thermal sensor-based automatic fever detection system which has the capability to scan a large number of people from distances up to 20 m and reduction takes place in real time where a person doesn’t even need to stop, compromising on the accuracy of the temperature measurement. This dramatically increases the usability of such system at places like airports, railway stations, metros, supermarkets, hospitals, exam centres, voting centres and other places where there is a large public movement involved.

This system is already currently in use in multiple co-working spaces and IT parks, and is going to be installed in other such locations including public transit and Hospitals.

It comes as a ready to use plug and play kit which can be set up in less than 10 minutes, and can process up to hundred people per minute. If anyone with the elevated body temperature is found an audio and visual alert is issued after which a manual verification can be done through a thermal meter and further action can be taken. This process is far faster than anything currently in use including industrial grade systems at airports. 

The system uses AI to detect faces and then measure temperature of the actual human from the face, rather than getting false alarms from other hot things like a cup of coffee/tea. 

Overview

  1. Thermal Camera converts IR radiations into detectable pixel values. Hence any objects temperature can be measured at a distance
  2. One of the first symptoms of Coronavirus is a fever, which results in a higher than normal body temperature
  3. Thermal cameras can be used for a preliminary screening of people having fever to eliminate chances of spreading the infection

Advantages

High Speed

A thermal camera can be setup in a fast moving space to automatically raise alerts if anyone with an elevated body temperature is found. Handheld thermal cameras take at least a few seconds to produce result and require a human to point it towards  a person

Monitor Crowds

By setting up the camera in a place, multiple people can be scanned at the same time without pointing the camera at anyone in particular. No congestion will be made by using this sensor

Improved Safety

Since the process is non-contact based, the risk of infection spreading due to contact is exponentially low

How does it work?

  1. Setup a channel for people to walk into in an orderly fashion (Such as a DFMD machine, turnstile, or walkways)
  2. Setup the camera at the termination of that line, so that all the people walking through will be scanned. Scanning each person takes < 1 second so line will be fast moving, and customer experience will not be disrupted)
  3. Camera will beep an alert for anyone with high body temperature who can be pulled aside and manual thermometer verification can be done

Features

  1. Low setup time
  2. Measurement from ~20m
  3. Real-time detection and alerts
  4. 80-100 persons / minute throughput
  5. Custom enclosure to be discrete