In this paper we have presented a real-time algorithm to track and recognise hand gestures for humancomputer
interaction within the context of videogames. We have proposed an algorithm based on hand
segmentation, hand tracking and gesture recognition from extracted hand features. The system’s performance
evaluation results have shown that this low-cost interface can be used by the users to substitute traditional
interaction metaphors. The experiments have confirmed that continuous training of the users results in higher
skills and, thus, better performances.