Researchers from MIT and the University of Minnesota have developed a next-generation security surveillance system capable of picking out signs of trouble from security video feeds faster and more accurately than human can.
The surveillance system works via artificial intelligence to ‘remember’ patterns from recorded footage; scan real-time feeds and identify potential signs of trouble or specific suspects. The system can also identify unusual, potentially risky activities in places such as bags deliberately left unattended at airports.
It eliminates potential pitfalls in security surveillance such as the inability of human monitors to accurately scan through numerous video footages and identify suspects’ faces in record time. The new system is capable of raising the alarm over potential danger in record time, while preventing false alarms.
According to Christopher Amato, a postdoc at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), the system also factors context when analysing a set of images. For instance, in an airport, the system could be programmed to identify and track certain individuals or pick out objects or strange things in unusual places. It works by identifying unusual patterns, such as objects moving in unusual ways or strange places.
Follow Computerworld Australia on Twitter: @ComputerworldAU