Human computer interaction for the disabled with upper limbs amputation

A robust human-computer interaction is proposed. Through robust facial features location and tracking, the disabled with upper limbs motion impairments, e.g. amputation can easily operate computer. It has the following contributions. (1) The work is based on common camera instead of infrared device. Therefore, the cost is low. (2) The enhancement facial motion detection based on local area is presented to simulate the mouse click events. (3) A trajectory smoothing algorithm is proposed, which can improve the accuracy and stability of the facial motion tracking. Experiments show it can robustly track the facial movement under the background and illumination variation. It can also effectively detect the facial expressions.