Classification of Renal Stone Types In Ultrasound Images Using Texture Analysis

Abstract

This analytical cross-sectional study aim to classify renal stone type using Ultrasonography. The density of renal stone usually determined the type of treatment where the stone density ≤ 1000 HU (type I) the patient need Extracorporeal Shock wave lithotripsy and if the stone density ≥ 1000HU patient need surgery and the density can be obtained through Computed Tomography. The general objective of this study was to reduce the radiation dose to patient by using Ultrasonography, Computed Tomography image cost and time. The data was collected using data sheets from three centers (university of kordufan diagnostic centre, Eldaman hospital and Elsalama clinic centre), The study has been carried out during the period from April 2017 up to October 2019. The patient data were statically analyzed by SPSS and M.S Excel. The stone density was measured to 90 patients’ 68 males (75.6%) and 12 females (13.3%) the mean stone density was found to be 808HU, the mean of stone length was found 1.53cm, the mean stone width was found 1.08cm, while the mean age of patient was 49 years and the mean stone area was 2.24cm2. 61 patients stone density type I and 29 patient type II. then the some data were collected from those patient using ultrasound images and by analyzing images by IDL in order to classify the stone into type I and II using linear demonstrate analysis and texture feature as input data, the result of classification showed that the overall accuracy was 96.7% and for stone type one was 100% and type two was 91.7%, from this program the consultant could chose the best decision for the patient.