Gesture Recognition Method Based on Fourier Descriptor
This paper proposes a new gesture recognition method based on fourier descriptor to solve the problems of palm contour feature extraction and low gesture recognition rate.Firstly,this paper segments the palm region based on the maximum contour fourier descriptor matching and confidence of the candidate window.Then,this paper tracks the palm contour and calculates its fourier descriptor to obtain the gesture contour feature values.Finally,this paper utilizes the BP artificial neural network to receive the 16 gesture contour feature values and recognize the gesture categories.Experiments have shown that the proposed method can effectively capture gesture contours and recognize 19 categories of gestures,with advantages such as high recognition rate,excellent performance,and good robustness.
NETL
NSTL
万方数据
