A multi-segment image coding and transmission scheme

In this paper, we present a multi-segment image coding scheme suitable for lossless progressive transmission over noisy or congested networks. The method is a block-wise forward-adaptive context selection technique that employs a low-complexity classification process to separate an image into multiple segments according to local statistics, so that compression efficiency by an entropy coder is improved and progressive transmission becomes possible. At the same time, loss of information in case of error occurrence during transmission results in less obvious quality degradation as the lost blocks are scattered in the final image. The scheme was extensively tested and turned out to give results comparable to those given by state-of-the-art standards, such as L-JPEG, JPEG-LS, CALIC and FELICS, in terms of compression efficiency, while providing with scalability and error concealment mechanisms. (C) 2005 Elsevier B.V. All rights reserved.