Research And Application Of Depth Information Based Gesture Recognition

Human-Computer Interaction technology is an important research field of computer science, in recent years body interaction is a hotspot in the research and application of Human-Computer Interaction. This article selects gesture recognition based on depth information as a research direction and uses the recognized gesture to drive the application of scene in the three dimensional virtual library.Traditional gesture recognition mainly has two methods of based on data gloves and based on vision. These two methods have some limitations in the application. The method based on data glove has good interaction effect, but users need to wear special sensor gloves. This device makes the way of interaction not be natural and reduce users’experience. The method based on vision is affected by skin color and the light of background easily, leading the recognition rate is low. This article recognizes fingers based on depth information and Space Geodesic Shortest Path, tracks finger motion using particle filter, then determines gesture according number and name of recognized fingers. Through the experiment, results verify the effectiveness of finger and gesture recognition.CADAL digital library is a good online library. As a web portal, it provides book borrowing, reading, personalized recommendation and other services. It makes great contribution for the spread of knowledge and protection of natural resources. With the development of virtual reality and HCI, three-dimension scene is used more and more widely. Based on the digit resource in the library, this article builds a virtual library, gives the design of general framework and functional module; According to the structure of virtual library, this article gives solutions of book retrieval, collision detection and people interaction, realizes each function module of the scene. Finally, we use the recognized gesture to drive information of scene and realize body interaction in the scene.