As economies gradually reopen globally, people around the world share a responsibility to keep up the fight against COVID-19. To allow more businesses and people to monitor social distancing easily, AI Hub Singapore has created an AI application that turns a ubiquitous device – a mobile phone – into a social distance monitoring instrument.
Dubbed SafeDistancer, the innovative application makes use of AI and computer vision to detect people in the phone’s camera frame and monitors the distance between them in realtime. When people get too close together, the application is able to emit alarms to highlight the breaches.
“Our customers and partners are looking forward to going back to their workplaces and we are happy to be able to help them do this in a safe and responsible way,” said Chong Choo, Director of Ecosystem at AI Hub Singapore.
The “AI at the edge” system runs advanced machine learning algorithms to detect people in an image. The results are then processed with data analytics on-device to alert when people get too near each other.
A key concern in such computer vision AI systems is the need to respect personal privacy. SafeDistancer addresses this in two ways. First, all of the processing is done on-device so there is no sending of images across a network – the mobile phone can be used without network connectivity. Second, the system is designed to not recognize faces and also does not store any images on the device.
SafeDistancer is currently being rolled out with key partners as businesses get ready to resume normal operations. In addition to running on a mobile phone, it also comes with the ability to be integrated with existing enterprise infrastructure.