Multiple description image coding using pixel interleaving and wavelet transform

Multiple description coding encodes a source image into two or more mutually refinable bitstreams to overcome channel impairments. When one or more channels fail, received bitstream(s) are decoded to provide a lower but acceptable quality of the reconstructed image. This paper proposes a multiple description image coding scheme using pixel interleaving by taking advantage of the intrinsic correlation of the adjacent pixels. We improve substantially the performance of the reconstructed image from multiple descriptions by compromising to some extent the performance of the reconstructed from a single description. This is achieved by incorporating the wavelet transform into the pixel interleaving-based encoder so as to provide the ability to trade off the quantizations of the frequency coefficients and the spatial pixels. Simulations are presented to demonstrate the effectiveness of the proposed schemes.