Facial recognition used to streamline roll call in Brazil schools
The school roll call as most know it is being transformed by face biometrics. In Paraná, Brazil, over 200,000 students were processed daily during the pilot phase of a new facial recognition-based attendance system developed by Valid and Innovatrics. It has cut down roll call time by as much as 80 percent, according to a case study.
Innovatrics’ compact neural network for direct face detection on devices authenticates identity by processing kilobytes of data instead of transmitting an entire video stream that could be several megabytes. Teachers can now use a mobile app on their phone developed by the government and take one to four pictures of an entire class.
The detection software extracts each students’ face from the photos. Average face detection takes one millisecond and face extraction averages 100 milliseconds. Faces are compared to photos in a watchlist. The software marks identified students present or absent. Names left unticked are manually checked by teacher. Parents also have the app.
The program eliminates the need for traditional physical hardware, which can have access-control benefits. It also reportedly has zero hardware maintenance costs, while the traditional system used 17,000 servers worth over $500,000. The system has the potential to be implemented in more than 2,100 schools, affecting almost one million students.