Palm Gesture Recognition Technology Based on Comparison of Contour Autocorrelation Functions
The technology of detecting static hand gestures is considered. The technology includes four main stages of image processing: detection of palm contours; piecewise linear approximation of the palm contour; construction of the autocorrelation function (ACF) of the contour of the observed gesture; comparison of the calculated ACF with the ACF of reference gestures. Examples of identification of the main gestures with the palm using the proposed technology, as well as the resulting quantitative estimates of the similarity measure of gestures, are given. The proposed technology is characterized by high speed and lack of learning process. The presented approaches may be of interest in the construction of non-contact control systems.
static gesture, edge detector, piecewise linear approximation, autocorrelation function