Classification of urinary stones based on edge detection using semi-automatic threshold

Abstract

Urinary stones are commonly differentiated into four types; calcium, struvite, cystine, and uric acid. While two of them, cysteine and struvite, could not be distinguished easily, in fact, they need a different kind of treatments. Significantly, the purpose of this study was to classify of urinary stones based on edge detection using a semi-automatic threshold. Five urinary stones were scanned using eight sources voltage. The source voltage was 65, 75, …, 115 KV. The consistent parameters were set at; 0.20 rotation, 13.89 µm resolution, and 1 mm aluminum. The reconstructed images were quantified in CT Analyser 1.16.4.1. The region of interest (ROI) was a circle with a diameter of 275 pixels as well as the semi-automatic threshold were drawn on the reconstructed image. Based on the Hounsfield unit (HU) analysis, there were three groups of urinary stone on source voltage of 85 KV. The first group (struvite) has consisted of stone number 1 (461.42) and 3 (489.05). The second group (cystine) has consisted of stone number 2 (55.08) and 5 (-233.41). The third group was stone number 4 (634.18). The edge detection using semi-automatic threshold can classify five urinary stones into three types; struvite, cystine, and the other group.