The explosion in Internet use has led to a growing demand for video downloading and streaming services via a wide range of bandwidth and devices of varying storage and processing capacity. Perceived quality of these services, in most of the cases, is not up to the expectation level.
This project investigates to improve video quality by exploiting block-level moving regions with adaptive pattern templates and handling object-occlusion problems with dynamic background modelling. Results of this project will contribute in the next-generation video coding standard supporting a wide variety of devices from mobile phones to HDTV, using wired/wireless technology, and covering the Internet as well as terrestrial/satellite broadcasting networks.