Efficient scalable DCT-based video coding

It its well-known that flexibility and error resilience are significantly improved by employing a scalable bit stream. The major drawback of multi-layered representations within a motion compensated (MC) discrete cosine transform (DCT) based framework is the increase in bit rate as compared to a single-layered representation having the same frequency, spatial and temporal resolution as in the highest layer of the multi-layered representation. This increase in bit rate is due to side information overhead, variable-length coding inefficiencies, and the differing statistics of the error signal. Consequently, much of the research in the area of scalability has focused on non MC-DCT based techniques having inherently scalable properties, e.g. sub-band techniques. However, the ubiquity of MC-DCT based technology suggests that we also address the problem within the MC-DCT framework. This is further warranted given the inclusion of syntax extensions to support scalable coding within newer MC-DCT based video coding standards. In this paper we present a rate-distortion optimized SNR and spatially scalable framework for MC-DCT based video coding.