Real Time Interactive Presentation Apparatus based on Depth Image Recognition

Abstract

The research on human computer interaction. Human already thinking to overcome the way of interaction towards natural interaction. Kinect is one of the tools that able to provide user with Natural User Interface (NUI). It has capability to track hand gesture and interpret their action according to the depth data stream. The human hand is tracked in point of cloud form and synchronized simultaneously.The method is started by collecting the depth image to be analyzed by random decision forest algorithm. The algorithm will choose set of thresholds and features split, then provide the information of body skeleton. In this project, hand gesture is divided into several actions such as: waiving to right or left toward head position then it will interpret as next or previous slide. The waiving is measured in degree value towards head as center point. Moreover, pushing action will trigger to open new pop up window of specific slide that contain more detailed information. The result of implementations is quite fascinating, user can control the PowerPoint and event able to design the presentation form in different ways. Furthermore, we also present a new way of presentation by presenting WPF form that connected to database for dynamic presentation tools