Adaptive Key Frame Rate Allocation for Distributed Video Coding

In the context of Distributed Video Coding (DVC), rate allocation among sources represents an important problem to solve. While in the information theoretical setting of Distributed Source Coding (DSC) the statistical correlation between sources is usually assumed to be known, in practical DVC systems there is no a priori knowledge of the underlying statistics of visual data. This lack of information makes it difficult to deal with the problem of rate allocation in practical DVC codecs. In this paper we focus on the problem of how to distribute the rate between differently encoded parts of the video sequence in a DVC system. Namely, we propose an adaptive rate allocation scheme for the encoding of the key frames depending on an estimation of the local motion activity of the sequence.