Development of Technology for the Identification of Model Parameters for Dendritic Structures Images

摘要

This work aims to increase the reliability of dendritic crystallogram's images classification. Crystallographic methods are used for medical diagnosis and we propose here to improve the reliability of their classification through an improved description of de dendritic structures' features. In this paper, we use the parameters of the mathematical model describing objects with dendritic structure. We developed a technology of parameters identification from a model image of dendritic structures, that was then implemented through the use of geometric and statistical features, together with a nearest neighbor classification algorithm.