ThumbStick: a novel virtual hand gesture interface

Nowadays most of the gesture interfaces need two hands to give user input. However, they are not user-friendly for the mobile users and disables. In this case, there is a need for new user input methods such as using one hand. In this paper, we present a framework of video-based virtual hand input for using one hand called ThumbStick, which consists of six key components in hand analysis & tracking - skin color segmentation, noise reduction, edge detection, connected component labeling, a contour line condensation tracker with adaptive searching path, and thumb movement recognizer. Unlike the previous systems, this system not only allows user to use one hand to give input, but also provide five degrees of freedom.