Research On Gesture Recognition Based On Zernike Moments And Support Vector Machine

With the development of computer technology and sensor equipment,human computer interaction is becoming more diversified.Natural human computer interaction technology is that user exchanges with computer through the voice,action,gestures,expressions and other natural ways.Gesture recognition is one of the important components of natural human-computer interaction technology,which has important research value and wide application prospect.This paper focus on the study of vision based hand gesture recognition using Kinect v2 sensor.In terms of gesture acquisition,Kinect v2's skeletal tracking technique is used to determine the hand area.With the depth of the data,the hand is divided from the background,and the salt and pepper noise is eliminated by median filtering.In the aspect of dynamic gesture tracking,the traditional temporal templates method was improved,the change of position of the hand in the vertical direction is added to the model parameters of motion gestures.Through the analysis and comparison of Hu moments,Zernike moments and Wavelet moments,the Zernike moments is selected as feature extraction method.Combined with the characteristics of the data samples,the support vector machine is selected as the classifier.The combination of Zernike moments and SVM has achieved good results in the identification of dynamic gestures and static gestures.Finally,a WordPad application based on gesture recognition is proposed,which can achieve multiple functions such as writing,creating,saving and so on.There are some innovations for researches in gesture recognition which has significant application innovation value.The work in this article is also a good base of research and development works afterwards.