There are presented here a concept and the results of testing of a method of textures recognition based on calculation of serial statitics reached from sequences of pixel values drawn from histological images. Serial statistics make possible a discrimination between pairs of samples drawn from different statistical ensembles when their probability distributions are a priori unknown. The concept of such statistical tests has been here adopted to the specific requirements of random fields describing the textures. An influence of the method of image scanning on the effectiveness of textures discrimination has been investigated. The problem of texture elements collection into segments covered by homogenous texture is also considered. Some natural constraints of statistical textures discrimination as a consequence of a "principle of uncertainty" has been indicated. The consideration have been illustrated by numerical examples.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.