The Applied Research Of Vision-based Gesture Recognition On Smart TV

The vision-based gesture recognition is a new and indispensable technology ofachieving new generation of human-computer interaction. Its application on the smartTV makes people get rid of the bondage of traditional Remote control andmanipulates the TV much more naturally and flexibly.From the perspective of natual human-computer interaction,this paper mainlystudies the application based on visual gesture recognition technology on the Androidsmart TV. First of all,from the development of the smart TV, it analyzes the style ofinteraction and reviews the development status of gesture recognition at home andabroad. Then it introduces system framework, the four components and Linux inputsubsystem of Android platform that the Smart TV used and gesture recognitiontheoretical knowledge. Meanwhile, it states the principle of how to achieve thegesture recognition on the Android platform. At the same time, through developingthe adapter in the Android application layer,it integrates the EyeSight gesturerecognition technology into the smart TV system.Therefore, it completes theapplication and achievement of basing on2D camera gestures recognition on Androidsmart TV. In addition, it analyzes the application of the3D gesture recognition on thesmart TV and the achievement of basing on3D camera gesture tracking and skeletontracking, especially the impact of the somatosensory games as immersive experience,reflects the treads of human-computer interaction from the abstract to the image andfrom low-dimensional information to the high-dimensional information, realizes theman-machine "natural interaction", adapts to the natural interaction guidelines of"people-centered" and creates a new natural interactive ecosystem.