Quality scalable low delay video coding using leaky base layer prediction

In this paper quality scalable video coding using leaky base layer prediction for low-delay applications is considered. The temporal prediction reference for the base layer is generated by calculating a weighted average of the quality base layer and quality enhancement layer reference pictures. This provides close to single layer performance at the enhancement layer rate point while introducing drift into the base layer if the enhancement layer is truncated or lost. Approaches using global weighting and locally adaptive weighting are investigated. The paper proposes an alternative approach for SNR scalable video coding targeted at applications usually close to the enhancement layer rate as operating point while providing good error recovery if enhancement layer packets are lost.