A new signage category that has emerged since the onset of the COVID-19 pandemic is “temperature-taking signage.” Last summer, I wrote about TempDefend, a “Protection as a Service (PaaS)” system from 22Miles. Another player that has appeared in this space is Altoros, a 20-year-old consultancy and professional IT services and software provider. Altoros has launched Mass Fever Screener, a combination of artificial intelligence (AI) cameras and Altoros-developed software designed to facilitate prevention of the COVID-19. The thermal camera scans people entering a building or other location (up to 30 people can be screened simultaneously), determines their temperature (with an accuracy of ±0.3°C/±0.6°F), and, if they are above an acceptable level, the system will alert security or whoever has been designated the responsible party, who then implements whatever policies specify what is to be done with an employee, customer, or visitor who has been found to have an elevated temperature.

One of the gaps in the market for temperature-screening systems that Altoros saw was that, said Andrei Yurkevich, CTO for Altoros, “no system that can orchestrate several cameras or are installed in multiple entrances.” Another gap in the market is data collection, which Mass Fever Screener integrates. “Where we’ve seen a need is in developing a system where it will be possible to collect historical data about people who enter the building,” said Yurkevich. “Most likely there will be some compliance requirements for organizations to track their employees or their visitors.” And it may be necessary to track statistics over time. “There was no tool to collect the statistics in an automated way,” added Yurkevich, “and to see how the numbers changed over time. How many people with elevated temperatures did you have last month, last week, now? The core of the system is reporting the details about how safe the environment is.”

Altoros started with the thermal cameras and the software, but as they saw additional needs arise, they started adding additional modules, including one that provides UV-based disinfecting. Another module helps enforce mask compliance. That is, the AI-based camera can determine if people who enter the location are wearing a mask or not, and, if not, trigger whatever action has been programmed (alert security, etc.). The system can also determine if the mask is being worn correctly (i.e., above the nostrils). What about a mask that is itself a face? For example, WhatTheyThink contributor Mary Schilling printed masks with the faces of WTT’s Adam Dewitz and Eric Vessels on them. Could the system tell the difference between a real face and a printed face?