Highly scalable wavelet-based video codec for very low bit-rateenvironment

摘要

We introduce a highly scalable video compression system for verynlow bit-rate videoconferencing and telephony applications around 10-30nkbits/s. The video codec first performs a motion-compensatednthree-dimensional (3-D) wavelet (packet) decomposition of a group ofnvideo frames, and then encodes the important wavelet coefficients usingna new data structure called tri-zerotrees (TRI-ZTR). Together, thenproposed video coding framework forms an extension of the original zerontree idea of Shapiro (1992) for still image compression. In addition, wenalso incorporate a high degree of video scalability into the codec byncombining the layered/progressive coding strategy with the concept ofnembedded resolution block coding. With scalable algorithms, only onenoriginal compressed video bit stream is generated. Different subsets ofnthe bit stream can then be selected at the decoder to support anmultitude of display specifications such as bit rate, quality level,nspatial resolution, frame rate, decoding hardware complexity, andnend-to-end coding delay. The proposed video codec also allows precisenbit rate control at both the encoder and decoder, and this can benachieved independently of the other video scaling parameters. Such anscheme is very useful for both constant and variable bit ratentransmission over mobile communication channels, as well as videondistribution over heterogeneous multicast networks. Finally, ournsimulations demonstrated comparable objective and subjective performancenwhen compared to the ITU-T H.263 video coding standard, while providingnboth multirate and multiresolution video scalability