Distributed video coding performances strongly depend on the side information quality, built at the decoder. In multi-view schemes, correlations in both time and view directions are exploited, obtaining in general two estimations that need to be merged. This step, called fusion, greatly affects the performance of the coding scheme; however, the existing methods do not achieve acceptable performances in all cases, especially when one of the estimations is not of good quality, since in this case they are not able to discard it. This paper provides a detailed review of existing fusion methods between temporal and inter-view side information, and proposes new promising techniques. Experimental results show that these methods have good performances in a variety of configurations.
展开▼
