Algorithms for Image Processing of Micro-Objects of Medical Diagnostics Systems
Abstract
Constructive approaches, principles, models for optimizing the identification of micro-objects based on the use of statistical, dynamic models with mechanisms for filtering noise, foreign particles on images of medical objects and pollen grains have been developed. Methods for optimizing the identification of micro-objects with the selection of characteristic objects in the image, the combination of image models with the mechanisms of segmentation, filtering, transformation, and alignment of the laser image processing zone with the CAD model are proposed. Algorithms for selecting the contours of objects on an image using a border detector in the transverse direction to each vector of the CAD model have been developed. Modeling of images of microsamples for medical diagnostics of chest diseases was carried out. The width, straightness of the boundaries of structural objects, local defects, parameters of regulation and synchronization of the components of the vision system are determined. Algorithms for image processing under conditions of a priori insufficiency, uncertainty of parameters, and low accuracy of data processing are studied. A mechanism for suppressing impulse noise and noise is implemented based on various filtering methods, preserving the boundaries of objects and small-sized parts. Mathematical expressions for estimating identification errors due to non-stationarity, inadequacy of approximation, interpolation, extrapolation of the image contourare obtained, which are synthesized with cubic, biquadratic, interpolation spline functions and wavelet transforms. A software package for recognition and classification of micro-objects has been developed.