Five students at the Santa Catarina State University in Brazil submitted a project to Computer Vision and Pattern Recognition subjects that functions as a pattern recognition system in a dashboard camera to detect the use of mobile phones of drivers while on the road.
The head proponent Rafael Berri, together with his friends, Alexandre Silva, Rafael Parpinelli, Elaine Girardi and Rangel Arthur introduced a solution to prevent drivers from using their mobile phones while driving. The camera monitors the pattern of the face, ears, and hands and other signs of ‘on the phone’ to lessen the chances of accidents.
Berri and his team said that drivers tend to fix their gaze straight ahead while on the phone. Thus, placing the dashboard camera in front of the driver is an ideal spot to effectively scan the patterns, according to MIT Technology Review. If the driver is caught while on the phone, the system sends a warning.
The team has tested the algorithm wherein the camera takes a set of five videos of the driver, 15 frames per second with 320 x 240 pixels in real-time. However, though the project worked in general – according to the team, “Periods of three seconds were correctly classified at 87.43 per cent of cases,” – external factors such as sunlight affects its accuracy when scanning the driver’s skin.
While the concept is possibly applicable, it is still questionable whether it would really prevent drivers from using their phones since hands-free driving devices are available in the market.