Presentation interface based on gesture and voice recognition

Abstract

In this paper, we introduce a Kinect based interface that recognizes gestures and voice. We have developed an interface to control presentations such as speeches or lectures. It is possible to receive the coordinates of the body, and recognize gestures and positions of the hand. Data received by the camera in Kinect are used to create a hook between the user hand and a presentation application such as Microsoft Powerpoint. Our interface is able to recognize grip and push gestures from the presenter. The result of this gesture recognition generates a signal to the presentation application, such as shortcuts to change slides or make use of additional tools. It is also possible to start and end the presentation by voice using our voice recognition tool. Additionally we show some tools that not only change the slides, but also provide more options to the presenter such as memo tools to directly highlight some parts of a slide, and even an eraser. This paper describes all the methodology and presents the result of our tests session. We are effectively able to improve the presentation capability of the presenter and think that such interface can be commercialized for presentation and other type of use.