Introduction: This paper shows the influence of a measurement method of features\r\nin the diagnosis of Hashimoto�s disease. Sensitivity of the algorithm to changes in\r\nthe parameters of the ROI, namely shift, resizing and rotation, has been presented.\r\nThe obtained results were also compared to the methods known from the literature\r\nin which decision trees or average gray level thresholding are used.\r\nMaterial: In the study, 288 images obtained from patients with Hashimoto�s disease\r\nand 236 images from healthy subjects have been analyzed. For each person, an\r\nultrasound examination of the left and right thyroid lobe in transverse and\r\nlongitudinal sections has been performed.\r\nMethod: With the use of the developed algorithm, a discriminant analysis has been\r\nconducted for the following five options: linear, diaglinear, quadratic, diagquadratic\r\nand mahalanobis. The left and right thyroid lobes have been analyzed both together\r\nand separately in transverse and longitudinal sections. In addition, the algorithm\r\nenabled to analyze specificity and sensitivity as well as the impact of sensitivity of\r\nROI shift, repositioning and rotation on the measured features.\r\nResults and summary: The analysis has shown that the highest accuracy was\r\nobtained for the longitudinal section (LD) with the method of linear, yielding\r\nsensitivity = 76%, specificity = 95% and accuracy ACC = 84%. The conducted\r\nsensitivity assessment confirms that changes in the position and size of the ROI have\r\nlittle effect on sensitivity and specificity. The analysis of all cases, that is, images of\r\nthe left and right thyroid lobes in transverse and longitudinal sections, has shown\r\nspecificity ranging from 60% to 95% and sensitivity from 62% to 89%. Additionally, it\r\nwas shown that the value of ACC for the method using decision trees as a classifier is\r\nequal to 84% for the analyzed data. Thresholding of average brightness of the ROI\r\ngave ACC equal to 76%.
Loading....