Vision-based Gesture Recognition And Its Interactive Application Research

With the development of science and technology,human-computer interaction is also developing in a more natural and humanized way.The traditional interactive methods can't meet people's needs.In recent years,augmented reality and virtual reality technology have developed rapidly,and further promote development of interactive technology.In addition,gesture recognition is widely used in many fields such as unmanned aerial vehicle control,smart home and sign language recognition.In this context,this paper studies the gesture recognition algorithm,and finally simulates the mouse function,and realizes human-computer interaction of natural human hand under monocular vision.The interactive system is composed of gesture segmentation,gesture tracking,gesture recognition,system realization.In the gesture segmentation,the skin color segmentation method for fixed threshold can't adapt to the real complex environment.The online skin color model is proposed,which is carried out in the follow-up gesture segmentation.Experimental results show that gestures can be effectively segmented.In the hand gesture tracking,the kernelized correlation filters is used to trace the gesture target.Aiming at the problem that the tracker is prone to track missing objects,two kinds of target re-detection mechanism were proposed.This paper uses the support vector machine and sliding window to detect the human hand,but the sliding window traverses the whole picture,which brings huge time cost.According to the characteristics of hand movement and the static background,the detection area is reduced to a quarter of the original one by using the improved inter-frame difference method which detect the motion area and reduce the detection range.In the gesture recognition,using the Fourier descriptor as the gesture feature,the k-NN classifier is selected to identify the static gesture.For dynamic gesture recognition,a more simple method of counting statistics is proposed.This algorithm can meet the real-time requirement of the system.System implementation calls the application program interface function to achieve the simulation of the mouse,and a dialog box program was developed by using of MFC,which display gesture recognition results intuitively.