In these scenarios, the child has his representation in the form of a graphic character (avatar) in a virtual word. This avatar is used to perform actions suitable to the current state of the scenario development. Its behavior is dependent on the child’s activity. The narrator informs the child which exercises/gestures he/she should do to achieve results which are expected at the given stage of the scenario. If the vision system recognizes one of these activities then the action in the scenario related to this activity is performed. Fig. 6 shows a typical scenario outline.