3D, 3-band, 3-tap temporal lifting for scalable video coding

Recently, a new generation of video compression techniques gained the interest of the research engineering community, by proposing state-of-the-art compression performance in addition to temporal/spatial/SNR scalability features. They exploit interframe redundancy through a temporal wavelet transform in the motion direction. While classical two-band decompositions offer dyadic factors of scalability, in this paper we present a three-band temporal lifting structure providing a factor 3 of scalability in a motion-compensated subband video coding scheme. More flexibility is thus available in order to adapt the bitstream to the bandwidth variations in heterogeneous networks. We also show improved compression performance compared with a dyadic Haar MCTF analysis.