Virtual Performance System Of The Gesture-driven Research

The vision based gesture recognition is an important part of new generation Human-Computer Interface. It makes the computer know the meaning of the hand gesture and response for. This will greatly change people's life. Now more and more people prefer to play music by themselves. The gesture recognition technology of man-machine interactive technology is gradually mature. It makes virtual music play possibility.By deeply research the gesture recognition technology of natural human-machine interactive, this paper proposed a way of combing the natural human-machine interactive technology and music performance to successfully resolve the Music-lover's vexation. This riches people's spiritual life, and thus has a certain practicality and scalability.This paper firstly compares several common computer gesture recognition technologies, and select the most suitable one the computer-visual based gesture driving technology as research priorities. Then, against the distance in gesture recognition process, this paper gives out the corresponding solutions for human-computer distance by Computer-vision based three-dimensional distance measurement, and gives the distance measurement formulas are derived. We need to go through several important stages to achieve the real-time user gesture recognition:gestures modeling, gesture segmentation and gesture analysis. Among them, gestures segmentation operation is the key part of this paper. This paper illuminates the technology to extract human region based on human color features dynamically, and further refined the feature data of the gestures segmentation model using Freeman chain principle,Finally, this paper realizes musical virtual performance system, which is based on gesture recognition technology. This system inputs gesture using single-eye camera and compiles Visual C++ console application program for musical performance. Finally, the paper shows the concrete realization of the system.