Abstract
The propagation of entropy rates in the formulation of Kolmogorov-Sinai (K-S) entropy forms the basis for identifying the dynamical behavior of a complex system often expressed in time-series data. This paper presents a procedure for estimating K-S entropy of images, which can be used to study the mechanism underlying the spatial content of images and useful as a multidimensional feature for pattern classification. The proposed method has been tested to detect spatial characteristics of different scenes and spatial objects. The potential applications of the proposed method appear to be promising for chaos analysis of images.