AIIR Scan Contactless Temperature Measurement by Alchera Inc. Helps Restaurants at San Pedro Square Reopen
By linking body temperatures with facial recognition data, Alchera’s contactless temperature screening system can minimize viral spread and help organizations open safely.
How do institutions with high foot traffic remain accessible and viable during the COVID-19 pandemic? That’s the challenge faced by schools, hospitals, places of worship, and commercial buildings — among others — as they seek out solutions that balance public safety, cost, and operational feasibility.
San Pedro Market Square, a multi-tenant building with a dozen foodservice outlets in downtown San Jose, CA, is a case in point. Closed for two months after the pandemic hit California, the property owner asked Alchera to use its artificial intelligence (AI) and access control expertise to develop a temperature scanning system that would help its foodservice tenants provide a safe dining environment for customers.
Using facial recognition algorithms to authenticate identity and control gate access is a specialized application of AI technology. Extending that capability by adding a layer of temperature data represented an opportunity to develop an innovative system that leverages Alchera’s in-house engineering and data science expertise.
Alchera proposed a two-phase temperature screening proof of concept (POC) plan. The proposal was accepted by San Pedro and work began in March 2020.
During Phase 1, the aim was to field-test a scalable temperature screening system for about 70 people — market employees and third-party vendors who regularly visit the shops. During Phase 2, scheduled for the coming months as the city and state work to reopen officially, the system would screen customers, too.
Implementation started with the creation of a client-side database. Each employee or contractor registered for the system by entering their basic contact information and taking a picture through a tablet or PC with a webcam. It takes about 60 seconds to create register each user; batch registrations are possible by uploading an Excel file. To ensure maximum data security, the database sits on the client’s server.
Next, the Alchera team set up a screening station at the San Pedro’s staff entrance. The station consisted of a simple kiosk with an embedded tablet and a side-mounted thermal camera.
The temperature screening system operates as follows. An employee walks past the entrance door and up to the tablet. The tablet screen prompts the employee to position their head inside an oval image on the screen. The oval helps to generate standardized temperature readings by positioning the employee close to the preferred distance from the thermal camera. The camera reads the temperature and scans the image for a match in the database. The whole process takes about 2 seconds.
The tablet screen displays results from the temperature scan. If the temperature is below a predetermined threshold (e.g. 101 degrees F), the tablet displays a green check that indicates a safe reading and the person is free to enter. If the thermal camera records a temperature over the threshold, the screen displays a red X. At the same time, the system automatically sends a message to the General Manager’s smartphone. Per San Pedro’s protocol, the employee is asked to leave the facility, get an appropriate health check or COVID-19 test, and return to work at the appropriate time. All information is saved to a web dashboard accessible by San Pedro management.
New social distancing tools will go into effect during Phase 2. These tools will allow the property owner to manage customer flow and site capacity (e.g. comply with seating limits) without incurring extra staffing costs. These tools include:
- temperature screening for unregistered users (e.g. customers)
- display a continuous count of users in the facility (e.g. track check-ins and check-outs)
Early lessons learned
Alchera’s AI can play a role in helping businesses and community organizations mitigate the social and financial costs associated with the COVID-19 pandemic. Based on our Phase 1 work, here are two lessons we’ve learned about implementation.
Appropriate screening technology choices depend on a checklist of site-specific factors. Large buildings (e.g. schools, stadia, commercial buildings, etc.) have unique layout features and constraints that need to be addressed during implementation design.
Clients need to balance temperature reading accuracy and budget. The accuracy of thermal cameras varies significantly. Typically, the higher the cost, the greater the accuracy.