Selective Inter−layer Prediction in Scalable Video Coding

In the scalable video coding (SVC) standard, spatial scalable coding outperforms simulcast coding when programs with several display resolutions are needed. Nevertheless, it is not suitable for end devices which only need the high resolution, due to a serious performance loss on the high spatial layer compared with single layer coding. To tackle this dilemma, a selective inter−layer prediction (SIP) method is presented in this paper. SIP attains an optimal trade−off by disabling inter−layer prediction on a set of selected frames. Theoretically, this selection can be modeled as a 0−1 knapsack problem which can be solved by dynamic programming. [continued on the next page]