A Shadow-Based Simulation Framework for Interaction Game

In this paper, we design and implement an effective framework for interaction simulation based on real shadows and virtual shadows. Within the specified depth range, we firstly identify different physical objects with the method of HSV And then in order to get the position of interaction tool, we use central moment to find the centroid of the physical object. In addition, the gesture recognition method is based on the shadow images by a web camera. Firstly, we set the region of interest on the screen. Secondly, we separate the hand shadows by the binary image. Finally, we use the convex hull, the defect and Hu's invariant moment methods to distinguish gestures and motions based on geometrical feature. The registered virtual shadows will create and interact according to the change of the real shadows. In the end, several special experiments show that our shadow-based simulation framework has well recognition and tracking efficiency.