Implementation of a robust absolute virtual head mouse combining face detection, template matching and optical flow algorithms

Abstract

This work proposes the implementation of a robust absolute virtual head mouse based on the interpretation of head movements and face gestures captured with a frontal camera. The procedure combines face detection, template matching and optical flow algorithms to emulate all mouse events. This virtual device is designed specifically as an alternative non-contact pointer for people with mobility impairments in the upper extremities. The implementation of the virtual mouse was compared with a standard mouse, a touchpad and a joystick. Validation results show motion performances comparable to those of a standard mouse and better than those of a joystick in addition to good performances when detecting face gestures to generate click events: 96% success in the case of opening the mouth and 68% in the case of voluntary eye blinks.

This work has been partially funded by Indra and Fundación Adecco under the Accessible Technologies Chairs Program