Avtor/Urednik     Brecelj, Janez
Naslov     Ocene dinamike bolezenskega procesa s statistiko prepoznavanja vzorca na rentgenogramih: analiza dejavnikov razvoja solitarne kostne ciste
Tip     monografija
Kraj izdaje     Ljubljana
Založnik     Medicinska fakulteta
Leto izdaje     1999
Obseg     str. 51
Jezik     slo
Abstrakt     The subject of the present thesis is a benign bone disease, the solitary juvenile bone cyst. The principle aim of the work is to study the influence of various factors on the dynamics of the disease. All three hypothesis have been confirmed. I. The first hypothesis assumes that it is possible to recognize the disease and evaluate its dynamics by the statistical analysis of X-ray images transparency For this purpose, 527 X-ray images of diseased bone in 48 patients have been analyzed. The images have been digitized with the MCID M4 Image Analyzer (Imaging Research Inc. Canada). Based on an anatomical point, three regions of interest (ROI) have been determined on each image; regions x and y through the affected part of the bone and region z through the healthy part of the bone. Image information encompassed in ROI has been translated into numerical values ( greys' value for each point). That values represents the statistical data suitable for quantitative analysis. The transparency of the X-ray images is quantified with coordinates of transparency. The distribution of the relative transparency of the affected part of the bone with respect to the transparency of the healthy part of the bone was evaluated. Coordinates of transparency were defined by the quantiles of these distribution function. The relative transparency with respect to the healthy part of the bone on the same image, could be compared among different patient-images, which enables statistical modeling. The complete statistical model includes a model of the disease status and a model of dynamics of the disease. The model of the disease status explains expert scores of the disease status with the coordinates of transparency. An empirical variable describing the disease status is constructed by this model. (Abstract truncated at 2000 characters)
Deskriptorji     BONE CYSTS
FEMUR
HUMERUS
METHYLPREDNISOLONE
CHILD
TREATMENT OUTCOME
REGRESSION ANALYSIS
SEX FACTORS
AGE FACTORS