For example,Бобцов

SPECTRAL CHARACTERISTICS OF CONTOURS OF LUNG SPHERICAL FORMATIONS

Annotation

A two-dimensional array of densitometric parameters obtained by processing tomographic medical images of patients with verified diagnoses of lung diseases (cancer and tuberculosis) is analyzed. The possibility of using information about the properties of contours of lung spherical formations is investigated. A technique for constructing contours of such formations based on the analysis of digitized medical images is applied. A mathematical method introduced for describing the contour is based on calculating the conditional center of mass of the image of the pathology and obtaining two one-dimensional signature functions, a module of the radius vector drawn from the center of mass and its polar angle, as functions of the conditional number of the pixel belonging to the contour. Using the spectral analysis of digitized medical images, such characteristics of the contours of the formations under investigation as the spectral entropy of the contour, multiplicity, and the number of local maxima, are calculated. Analysis of statistical difference of these characteristics for different types of diseases (cancer and tuberculosis) was carried out. It is shown that several studied characteristics of the contours are statistically different and therefore enable differential diagnosis of lung disease. 

Keywords

Постоянный URL

Articles in current issue

CONTENT © ITMO UNIVERSITY 1993-2024