I was traveling to New York city via JFK airport during the covid19 outbreak. The airport had passengers waiting for over 6-7 hours to be individually measured by TSA personnel to ensure travelers has no fever. To manually, individually measure a person’s temperature is simply unrealistic for future travelers at airports, hospitals, schools, businesses, etc.

What it does

The system is a set of thermal and video cameras with AI deep learning algorithms that can detect a person with fever from a distance (10-20 feet).

The system will capture the person’s face and compare it against a database for face recognition. The database of persons with fever will be logged and tracked to ensure they are isolated and tracked by public health professionals to prevent the person from spreading the virus.

How I built it

We use the company IronYun’s AI video analytics as a base with thermal cameras from Flir and other video surveillance companies.

Challenges I ran into

To be able to accurately detect a person’s skin temperature from a distance of 10-20 feet to within +/- 0.3 degrees Fahrenheit is very difficult. We had to invent new AI deep learning algorithms to learn what is the “abnormal” body temperature.

Accomplishments that I'm proud of

The system can help save lives and reduce the spread of COVID19.

What I learned

To develop a contactless system that can detect a person with fever from a distance is a very challenging problem.

What's next for Detection of person with fever with AI video analytics

To deploy the system in production and help reduce the spread of COVID19.

Built With

  • deeplearning
  • tensorflow
+ 1 more
Share this project: