Efficient layered video coding using data partitioning

摘要

Data partitioning has been incorporated within the MPEG 2 video coding standard as one means of achieving error resilience in transmission systems where more than one transmission priority level is available. However, data partitioning also includes many ideas from the study of frequency scalability. In particular, the base partition information could be used to form a lower spatial resolution reconstruction of the original service thus permitting service interworking. In this paper, we investigate service interworking using data partitioning and demonstrate that the quality of the lower resolution service when this approach is used is fundamentally limited by drift. It is also shown that a relatively low data rate correction signal can provide a significant increase in the quality of this lower resolution service.