Camera Control and Multimedia Interaction using Individual Object Recognition

Abstract— Currently, most of the automated, computervision assisted camera control policies are based on human events, such as the speaker gesture and position changes. In addition to these events, in this paper, we introduce a set of natural camera control and multimedia synchronization schemes based on the individual object interaction. We describe in detail, how our unique method, in which the head-pose estimation are used to compute the region of interest (ROI) for recognizing the hand-held object. We explain, from our results, how our approach has achieved robustness, efficiency and unambiguous object interaction during real-time video shooting.

• Type: Journal
• Year: 2007
• Where: JMM2
• Authors: Richard Y. D. Xu, Jesse S. Jin
• Keywords: Camera Control, Camera Control Policies, Individual Object Interaction, JMM2 2007