Stochastic Image Modeling Using Cumulants With Application To Predictive Image Coding
- 24 August 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 239-244
- https://doi.org/10.1109/hosa.1989.748771
Abstract
Stochastic image representation techniques traditionally rely upon second-orc[er image statistics to determine the model parameters. This paper reports on the use of third-order cumulant statistics to imp] ement non-causal, phase sensitive ARMA image models. A novel, linear equation based 2-D MA parameter estimation algorithm is extended from an existing 1-D algorithm, for modeling colored prediction-error residuals, and is used in conjunction n with a non-causal 2-D ARMA parameter estimation algorithm. A weighted Ieast-squmes MA approach is also developed as has been done in the 1-D case. Application of cumtdant-based stochastic image representations to predictive image co,iing is discussed and preliminary results using causal and noncausal ARMA predictors are presented.Keywords
This publication has 6 references indexed in Scilit:
- On The Use Of Second- And Higher-order Inverse StatisticsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2005
- Identification of nonminimum phase systems using higher order statisticsIEEE Transactions on Acoustics, Speech, and Signal Processing, 1989
- Bispectrum estimation: A digital signal processing frameworkProceedings of the IEEE, 1987
- Statistical model-based algorithms for image analysisProceedings of the IEEE, 1986
- The importance of phase in signalsProceedings of the IEEE, 1981
- Advances in mathematical models for image processingProceedings of the IEEE, 1981