| Summary: | Development in human-computer interactions (HCI) has increased over the last decades due to the widespread use of computing devices such as portable computers and smartphones. While the conventional keyboard and mouse are effective, their use can be limited, therefore the hand gesture is an alternative interaction method. The development of a hand gesture recognition system will certainly improve the existing HCI systems. In this study, a hand gesture recognition system was presented using commercially available continuous-wave (CW) radar combined with a classification algorithm based on k-nearest neighbours. The principal component analysis (PCA) is proposed to extract features and reduce dimensionality of hand gesture spectra. Performance validation based on the leave-one-out procedure shows that the proposed system has almost 100% classification accuracy. © 2021 IEEE
|
|---|