Research On Quantization-error-based Distributed Video Coding

With the high development of wireless multimedia technology, the traditional videocoding methods are no longer applicable. As a new video coding technology, DistributedVideo Coding (DVC) has attracted great attention. Based on the Slepian-Wolf andWyner-Ziv information theory, it adopts intra-frame encoding and inter-frame decoding,which transfers a large number of calculations from the encoder to the decoder. DVC ismore suitable for wireless networks than the traditional video coding.Based on the transform-domain DVC, there is inevitable quantization error inWyner-Ziv frame for the uniform quantization. It has large quantization error especiallywhen the quantization is coarse. The quality of decoded image will become poor. Thispaper proposes a transform-domain scheme to correct the quantization error inWyner-Ziv frame. In this scheme, the information of quantization error in Wyner-Zivframe is extracted at the encoder. Then it is encoded using Low Density Parity CheckCodes. At the decoder, the information is decoded using side information. So thequantization error is corrected and the quality of decoded image can be improved. Thesimulation results show that this scheme of the coarse quantization can reach to thesimilar PSNR of the accurate quantization in TDWZ, while there is42%to59%reduction in computation.