Algorithms of two‐dimensional image identification and restoration based on an extended stochastic semi‐causal image model
Hiroshi Mizutani, Sueo Sugimoto · Electronics and Communications in Japan (Part I Communications) · 1981
Abstract This paper is concerned with a new digital processing approach for restoration of additive‐noise‐corrupted images without using a priori statistical information concerning the autocovariance function of the original image field and the variance of noise. Based upon an extended stochastic semi‐causal image model, the method mainly contains two algorithms of image identification and estimation. The algorithm for image identification consists of maximum‐likelihood estimation of unknown parameters and determination of the optimum image model by an information criterion (AIC). Furthermore, recursive Kalman estimation techniques are applied to restore the noise‐corrupted images.