Parameter Estimation of CAR Models for Classifying Wood Boards
- 24 August 2005
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 1376-1379
- https://doi.org/10.1109/icsmc.1988.712957
Abstract
Two dimensional random field models have been widely applied to various image processing problems. The sample images for a specific application are usually characterized by parameters of the model equations. The model parameters have been estimated by the least squares method (LSE), the maximum likelihood method (MLE), or the robust method. In this paper, a new robust method is proposed to estimate the model parameters of the causal autoregressive models (CAR). Preliminary experiments with the generated images are performed to compare the estimation performances. The proposed method is also applied to extract features for the problems of classification of surface defects on wood boards.Keywords
This publication has 5 references indexed in Scilit:
- Hierarchical Classifier Design Using Mutual InformationPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1982
- Robust StatisticsPublished by Wiley ,1981
- Robust Estimation of Power SpectraJournal of the Royal Statistical Society Series B: Statistical Methodology, 1979
- The Influence Curve and its Role in Robust EstimationJournal of the American Statistical Association, 1974
- The Fitting of Power Series, Meaning Polynomials, Illustrated on Band-Spectroscopic DataTechnometrics, 1974