Research Of Natural User Interface Based On Hand Gesture Recognition

As computer, a necessary tool, becoming an indispensable part of people’s work and life, we further gradually the research of human-computer interaction. Natural User Interface (NUI) is new interaction interface since Manipulation User Interface, Command Line User Interface and Graphic User Interface. Its purpose is to find an interaction method of people-centric and in accordance with humans’characteristics. NUI has applied to many fields and it is the most popular interaction method which will play an important role in human-computer interaction.The paper is based on the analysis and summary of NUI and traditional hand gesture recognition algorithm. It discusses the hand detection technology of skin color and motion information and real-time target tracking method. In the process of the hands segmentation, we proposed an adaptive threshold skin color model establishing method which is fit for the current scene and segmentation effect is better. In the extraction of skin color and motion information, the GPU parallel computing technology is adopted to improve the operation speed of the algorithm. In the process of hand tracking, we apply an improved CamShift algorithm based on recent trajectories combining the particle filter algorithm to ensure the real-time tracking, which is able to get the hand region when the targets lose.After recognizing and locating the hands, the position is mapped to the corresponding gesture command. With the help of holographic equipment, we design a set of natural user interface. It provides the function of zoom, rotation and dragging. And it can be applied to the new product release, teaching, stage showing or visual display. The operator can control the virtual object to achieve the man-machine interaction effect.