Low Cost Hand Gesture Recognition System Design and Implementation

Abstract

The design and development of a low cost hand gesture recognition computer interface, using a standard laptop webcam is presented. The purpose of the system is to recognise both static and dynamic hand gestures from a user in real-time and perform basic macro instructions on the Windows operating system. The calibration and gesture recognition processes are discussed. The system is able to correctly classify 19 static hand gestures and recognise 6 dynamic hand gestures with greater than 95% accuracy. However, the system has a varying latency of between 50–500 milliseconds due to inefficient interfacing with the operating system. It was concluded that an optimised operating system interface would improve system performance and user experience dramatically.