Noncausal predictive lattice model for image compression

摘要

In this paper, we introduce a noncausal predictive lattice model (NCPLM) for image compression. Owing to its full plane support, the NCPLM is a more effective model in reducing redundancy for improved image compression as compared to causal predictive models. To avoid instability and non-realizable synthesis problems in association with a noncausal support, the binary pyramid decomposition is adopted. The applications of the NCPLM for implementing lossless and lossy image CODEC are presented. Simulation results indicate that the proposed NCPLM is effective as a predictive algorithm for image compression.