No abstract

This paper presents a new image coding algorithm for scalable and robust image transmission. The algorithm, termed block-based layered set partitioning in hierarchical tree (B-LSPIHT), has both SNR and resolution scalabilities, and its performance is resilient to transmission errors. In the algorithm, the number of layers and the rate and resolution associated with each layer can be specified before encoding. Starting from the base layer, B-LSPIHT encodes one layer at a time until the design of the top layer is completed. The transmission errors are localized using a block-based method. The bits allocated to each block are determined by a dynamic programming technique based on the ratedistortion function computed from the channel models. Numerical results show that the B-LSPIHT algorithm is effective for constructing layered image transmission systems over noisy channels.

We present a novel method for scalable and robust image transmission over ATM networks. The algorithm is based on a scalable set partitioning in hierarchical trees (SSPIHT) coding scheme, in which the encoded image bit streams for packeting and transmission are divided into a number of layers. By accumulating different numbers of layers in the decoder side, images with different resolutions and distortions can be reconstructed. Decoders having different capacities therefore can share the same encoder. In addition, these layers have different degrees of significance so that transmission errors in insignificant layers will not cause severe degradation in performance for image reconstruction. Finally, we describe three basic schemes for the transmission of SSPIHT-encoded cells over ATM channels. Simulation results show that the algorithm is effective for the applications where robust and scalable transmissions are desired. © 2002 Society of Photo-Optical Instrumentation Engineers.