Abstract
Gesture recognition in human-machine interaction is a popular subject of study, but it also includes some problems. The main motivation of the proposed study is developed a novel vector-based feature extraction method for gesture recognition in an unmanned aerial vehicle (UAV) control. In this paper, the use of discrete wavelet transform on the signals acquired by multiple sensors and then, a statistical feature extraction from this transformed signals is proposed for the person-independent gesture recognition. In this way, it is aimed to get the invariant feature space according to speed and magnitude of the movement in different time slices. The success of the proposed method in the problem of person-independent gesture recognition is experimentally demonstrated with the comparative experiments.