Adaptive VoD architecture for heterogeneous networks based on scalable wavelet video coding

In this paper, the VoD architecture for heterogeneous networks based on scalable wavelet video coding is proposed, with enhanced spatial scalability. The system consists of the clients and server (C/S). Motion estimation is performed on the resolution of the wavelet domain using hierarchical variable size block matching (DWT-HVSBM). All the vectors of different spatial resolutions are encoded into multiple-layer motion scalability. The sub-pixel accurate MCTF is applied in the temporal frames, and the three-dimensional wavelet coefficients are encoded in 3D EBCOT, which is the extension of JPEG2000 code streaming. The server can adapt to the end users' requirements to provide the service over a heterogeneous network that has users with different bandwidths. The experiments show the performance of the codec is comparable to MC_EZBC, and the video on demand network system can simultaneously support multiple heterogeneous users' demands.