Optimization of micro-object identification based on detection and correction of distorted image points

The scientific and methodological foundations have been developed for the optimal identification of micro-objects, the use of traditional Gaussian, median filtering, fast Fourier transform, wavelet transformations, shift transformations, mechanisms using redundancies of information structural components and their features, statistical, dynamic characteristics of image points. Mechanisms for identifying micro-objects that have advantages in reducing the complexity and laboriousness of structure analysis and information processing, identifying and segmentation of the image contour, using growth dynamics, visual differentiation, extracting internal features and properties, approximation, and interpretation of objects are proposed. A mechanism has been investigated and implemented that performs the following functions: aligns histology slices; finds contours of objects, a set of levels, thresholds, combines segmentation, conducts registrations, forms a search graph, performs approximations based on a wavelet, shear, and other transformations, determines parameters, performs color coding and visualization of micro-objects. The signs of chronic inflammation - the presence of giant cells - were assessed. A software package for visualization, recognition, classification of images of micro-objects has been developed, the implementation of which has been tested under conditions of a priori insufficiency, parametric uncertainty, and low reliability of the information.

Перевод пока недоступен