Temporal signal processing: gestural stroke recognition application on communication gesture analysis and recognition

摘要

The aim of this study was to design a new gestural translator for human-computer communication, based on basic movements that children with cerebral palsy (CP) are able to produce consistently. One of the main feature of these children is that their poor motor performance does not allow the spatial control required to manipulate the mouse or keyboard. Nonetheless, they show residual motor capacities that are typical enough to be used as control commands. Based on a kinematic study of simple lines and curves drawings, an automatic gesture recognition system has been developed that analyses and interprets in real- time hand trajectories. Motor patterns are then identified and translated as computer functions.