Multiview video coding using cuboid data compression

Despite user demand, broadcasting of events from multiple viewing angles and depths has not been popular due to limited interactivity and severe bandwidth and computational constraints. This project investigates a novel theoretical framework for multiview coding using volumetric data compression. This new coding framework will be used to develop a new multiview video coding technique that addresses limitations of the current state-of-the-art. Our technique differs in joint-exploitation of image correlation in temporal and view domains, low frame dependency, and explicit handling of occlusion. The results will benefit commercial television networks, telemedicine, distance education, video surveillance, and other video analytics applications.