350 rub
Journal Biomedical Radioelectronics №6 for 2009 г.
Article in number:
The Spectral Analysis of Segments of the Image of an Eyeground for a Quantitative Estimation of a Vascular Pathology
Authors:
A.N. Bregneva, R.A. Tomakova, S.A. Filist
Abstract:
Research objective: reception of quantitative characteristics of two-dimensional spectra of images of segments of the eyeground, allowing to spend their classification by classes "norm-pathology", and also to define severity level of a pathology and-or its kind. For calculation of the informative parameters describing the image, two-dimensional analysis Fourier has been used. With that end in view in research it is offered to segment the image of an eyeground on characteristic segments and by modelling of images of segments and their pathologies to receive qualitative and quantitative parameters of pathological deviations of the investigated image. For experimental researches two models (the simplified images) one of segments of the image of the eyeground which is switching on a disk of an optic nerve (DON) have been chosen. At researches black-and-white images of an eyeground and its models were used. On one of these models of microcirculation in norm, on another - are broken. For reception of quantitative parameters the method of amplitude quantization of modes by means of which the exploratory analysis of the image of amplitude two-dimensional frequency spectrum Fourier was carried out was used. As a result of the exploratory analysis of models of images of segments with a disk of an optic nerve without a pathology and with a pathology of microcirculation the dynamic range of quantization in which the segment image "was entered" has been defined. On the image of a two-dimensional amplitude spectrum of model of the image of a segment can be chosen set of the cells which histograms of brightness are displaced depending on a kind of a pathological deviation from the reference image. It has allowed as the quantitative characteristic of a deviation of the chosen segment of the image from norm to enter modes and amplitudes of modes of the chosen set of cells of the image, topology and which co-ordinates are defined as a result of the exploratory analysis. Thus, the offered method of synthesis of space of the informative parameters, a describing segment of the image of an eyeground with DON, allows to construct a complex of the solving rules based both on not fuzzy, and on fuzzy logic of decision-making, revealing pathological deviations in a vascular channel of an eyeground.
Pages: 15-18
References
  1. Ilyasova N.Yu., Ustinov A.V., Baranov V.G. An ExpertComputer System for Diagnosing Eye Diseases from RetinaImages // Optical Memory and Neural Networks. 2000.V. 9. N. 2. Р. 133-145.
  2. Патент № 2235496 (РФ).
  3. Грахов А.А.,Кузьмин А.А., Пихлап С.В., Филист С.А. использование метода адаптивного квантования мод при исследовании вейвлет-изображений электрокардиосигналов больных ишемической болезнью сердца // Известия ЮФУ. Технические науки. Тематический выпуск. «Медицинские информационные системы» (Таганрог). 2008. №5(82). С. 76-79.