A Robust Low Cost Virtual Mouse Based on Face Tracking

This paper introduces a non-contact virtual mouse controlled by tracking the face movement. It is specially designed for the disabled with mobility impairments in the upper extremities. The whole work is based on the common Web camera. Improved optical flow algorithm is proposed through locating the facial features defined by active appearance model. Experiments show it is robust to the various background and environment illumination. It also can effectively track the features when the face moves fast in a large scope.