UC 5.2  Vision -based human computer interaction application

Mobile handsets and tablets can activate, deactivate, or prevent their display from sleeping, depending on whether the user is attending to the screen. This use-case defines an approach for detection of user attention based on vision. A low-resolution camera is monitored continuously, and the presence of a forward-posed face is detected. Head poses looking off to the side (or up or down) are ignored by the system, as are conditions where no face is present.

This use case has both low-latency and low-energy requirements. Low latency is required to meet the goal of real-time user responsiveness. Low-energy is required as mobile handsets and portable electronic devices are uniformly battery-powered.

Synsense, in collaboration with UZH, CSEM and STGNB, works to reduce the battery or energy used by mobile handsets and tablets based on vision technique to know if the user is using or not the device. The device will accordingly move or not to the sleep mode. This solution will allow to increase the autonomy of the devices battery-powered and in general to reduce the energy used.