Two-dimensional optimal algorithms for image compression using ARMA predictors

Adaptive differential pulse code modulation has been extensively used in image data compression due to its relative implementational simplicity. It has also been demonstrated that the use of backward adaptive autoregressive moving average (ARMA) predictors in differential encoding of images lead to subjective improvements in the edge performance. Algorithms using constant convergence factors have been applied to adapt the coefficients of the adaptive predictor. Recently, two-dimensional algorithms using optimized convergence factors have been proposed and investigated. In this paper, two of these algorithms are used to determine the backward adaptive prediction filter coefficients. Simulations and comparative results using both autoregressive (AR) and ARMA predictors show significant objective and subjective improvements when the algorithms using optimum convergence factors are used