动态手势识别及在PPT播放当中的应用研究

本文主要研究了动态手势识别在PPT播放当中的应用,即在播放过程中能够通过动态手势控制PPT的翻页.首先在动态手势检测的研究中,通过采用背景差分和基于HSV的肤色建模相结合的方法,分割出了较好的动态手势图像.其次,通过对几种动态手势的跟踪算法的分析与比较,采用了主流的非线性跟踪算法W勇瞬’惴？通过对得到的手势图像中的V分量值的人为设置,消除了在跟踪过程中脸部及其他的类肤色干扰,从而避免了手势跟踪过程中的遮挡问题.最后通过采用隐马尔可夫模型与阈值Max相结合的方法,实现了对PPT播放的实时控制.实验结果表明动态手势的识别能够应用到PPT播放当中,并最终实现了利用动态手势来实时控制PPT的翻页.

Research on Dynamic Gesture Recognition and Its Application in Broadcasting PPT

This paper primary researches the application of dynamic gesture recognition in the broadcasting PPT, namely in the process of broadcasting through dynamic gesture to control PPT page turning. Firstly, in the study of dynamic gesture detection, this study by means of background difference combines with the HSV color space model method to realize dynamic hand gestures detection, the result indicates that can obtain better dynarn,ic gestures graphic. Secondly, this study compares several dynamic gesture tracking algorithms, then adopts the mainstream nonlinear tracking algorithm, that is particle filter algorithm. This algorithm eliminates face and other skin color interference in the tracking process, then this study changes the obtained values of the image V component, thereby avoids the gestures＇ cover problems in the process of tracking. Finally by using hidden Markov model combines with Max threshold value method, this study realizes the real-time control of PPT play. Experimental results show that the dynamic gesture recognition can be applied to the PPT of play, and finally realizes to control PPT page turning by using dynamic hand gestures.