A Multisacle based Multistage Lattice Vector Quantization Video Coding for Robust Transmission over Wireless Channels

In this paper a new non-predictive video codec for mobile applications is presented. The non-predictive scheme omits prediction step in the temporal axis. Hence, it increases robustness for transmission at low bit rate in mobile channel. The video sequence is grouped into a number of frames (GOP) and the same subband from each frame in the GOP are joined together to exploit their spatial-temporal redundancies. The significant vectors of the joined subband within the GOP are quantized using a novel multistage lattice vector quantization (MLVQ). This process reduces quantization errors and enhances the reconstruction quality of the frames. Experimental results of the new video codec are shown to be significantly better than H.263 and comparable to the current H.264 standard in erroneous Hilly Terrain mobile environment, simulated using TETRA channel simulator for some test video sequences.