Motion-compensated lossless video coding in the CALIC framework

In this paper we consider the problem of lossless compression of video sequences exploiting the temporal redundancy between subsequent frames. More specifically, we present a lossless video coding technique which extends interframe-CALlC by employing multi-frame motion compensation to first temporally decorrelate the video material, followed by context-based arithmetic coding of the residual data. To avoid the well known context-dilution problem, we perform context quantization. The proposed technique, called motion-CALIC, is shown to outperform competing methods such as interframe-CALIC and LOPT-3D up to 22.93% in terms of bitrate savings, while still maintaining manageable complexity