Recognition of hand gesture based on Gaussian Mixture Model

This paper presents a new method for gesture recognition of Human beingspsila hand. This method integrates the features of shape, color and orientation histograms, which are extracted from images, and estimate the comparability with all the different types of gestures by a proposed Expectation-Maximization algorithm in Gaussian mixture model. The classification results were presented based on the values of likelihood compared with all the types of pre-assigned images, and the performance of this approach in an experiment is shown that the proposed method works well.