A Multi-Touchpad based Human Computer Gesture Interface for People with Disabilities

Y. Yuan and K. Barner (USA)


Gesture Recognition, Human Computer Interaction, Neural Networks, Support Vector Machines


This paper describes a multi-touchpad based human com puter interaction (HCI) system called MGesture, which uses tactile gestures as input to perform common keyboard and mouse control. MGesture interface expands the abil ity of the multi-touchpad to incorporate an advanced tac tile gesture recognition engine and a fully customizable command processor. Users with physical disabilities can define their own gesture sets by providing training sam ples and associating them with corresponding computer commands. This application is equipped with two gesture recognition algorithms: recurrent neural networks and sup port vector machines based recognition. Preliminary exper iments show the effectiveness of the gesture recognition in terface and its usage within the multi-touchapd framework for users with disabilities.

Important Links:

Go Back