Current Studios worked with Clorox to develop an innovative beta device to help detect an uptick in illness indicators, including fevers, coughs and sneezes, in classrooms.
This beta device uses thermal analysis and machine learning to train algorithms to provide a source of vital, real-time information so that schools can make informed decisions around illness prevention. The device is placed at the front of the classroom, where it passively tracks temperature, coughs and sneezes. The software uses a custom machine learning algorithm pulling from a library of classroom noises to identify if the sound has the frequencies and amplitude of a cough or sneeze. When the algorithm identifies a cough or sneeze, the system logs it as a “hit.” The device does not record classroom audio. All audio is created and logged as a visual representation of the sound (in other words, a “spectrogram”).
Technology |
Machine Learning, Dataset Creation, Product Development |
Tags |
AI and Machine Learning |
.gif)
The device also passively track students’ temperature as they come within range to determine if they may be experiencing a fever (38°C/100.4°F). Thermal analysis detects the students presence and establishes a temperature reading by analyzing the thermal pixels at the inner part of the eye (tear duct) and if a reading reaches 100.4°F or above, it registers as an alert. This approach allows for a completely anonymous data set with no risk of tracking students. Instead it provides teachers and school administrators with the information to understand broaders trends and risks within the school.
Clorox and Dr. Saskia Popescu, an epidemiologist and infection preventionist, have developed methodology to determine a threshold for illness indicators, which we’ve incorporated into our notification system. When a temperature reading or the number of coughs and sneezes surpasses either threshold, the teacher and school administrators receive an alert via email. Educators can log into the dashboard at any time to see the data from that day, in addition to previous days and weeks. The end result is an ability to harness these illness indicators to help schools make informed decisions, while at the same time protecting the privacy of students and teachers.
