Virtual Interface And Its Application Research In Smart TV Interface

Human Computer Interaction is the process of information exchange, which is used to complete the task between people and computer in some way and in a certain language. The goal of human-machine interaction is to achieve the natural, harmonious and convenient interaction. With the development of computer technology, human-computer interaction technology is gradually shifting from the computer-centered to the people-centered. Gestures become a hot spot in the field of human-computer interaction as an important means of interacting with the outside world and expressing the emotion.Gesture interaction system based on vision mainly uses the camera to capture the human action and then processes information that the gestures express to complete the interaction tasks with the system. In gesture interaction system based on vision, because the hand is operating from a distance, it has the characteristics of non-contact and fuzziness. Therefore, computer processing system will indiscriminately process all gesture command within the scope of video. As a result, some unconscious gesture of the user can be interpreted as valid command, which causes system disorders. This phenomenon is also called "Midas Touch Problem" which is a common problem in gestures interface based on vision.This paper is supported by National Natural Science Foundation of China(No. 61472163), and the Science and technology project of Shandong Province(No. 2015GGX101025). This paper is aimed to solve the "Midas Touch Problem". The main research contents are as follows:(1) Put forward the concept of virtual interface, and use statistical analysis method to build cognitive basis of virtual interface.In order to explore the regularity of hand movement in the process of interaction, we design cognitive experiment. The experimental results show that although the different operators have different ways of operation and interaction, the operating region is mostly within a certain range, which is comfortable and convenient for the operator. The operating region is namely the virtual interface. In the process of interaction, gestures in the virtual interface are effective, and gestures outside the virtual interface are invalid. As a result, the virtual interface provides an effective method to solve the "Midas Touch Problem".(2) The algorithm of constructing the virtual interface is proposed and discussed.This paper puts forward two kinds of building algorithm of virtual interface. On the basis of the first construction algorithm, the second construction algorithm is presented considering the comfortableness of users. The second way to build virtual interface algorithm is divided into static construction and dynamic construction. Dynamic construction is that the operator draws a closed track in the air according to their own behavior, and the bounding box of the closed trajectory is the virtual interface. Static construction is that the operator raises his hand naturally and keeps the fist gesture. After a few seconds, the system will automatically set up a virtual interface centered the hand. The size of the virtual interface will be determined by the results of cognitive experiment.(3) Realize the partition operation and dynamic refresh function of the virtual interface.In order to complete different interaction functions through the virtual interface, partition operation of the virtual interface is put forward. Each region represents different functions, and the function distribution model is determined by means of questionnaire. Achieve dynamic refresh function of virtual interface, and it makes virtual interface can always exists in space no matter when user's location changes or in the case of user's memories forgotten, which makes users can interact with the system through the virtual interface.(4) In view of the relatively remote gesture recognition, a kind of gesture recognition method based on multi-region gesture distribution characteristics is put forward.Firstly establish two-dimensional rectangular coordinate system for Binary image to extract feature and then use the Euclidean distance and K neighbor algorithm to obtain the final recognition result. The experimental results show that this method can obtain satisfactory identification result.