Kidney Stone Detection using Butterfly Optimization Algorithm

Abstract

A kidney stone is a solid object made of molecules found in urine. Uric acid, cystine, struvite, and calcium oxalate are all possible causes of kidney stones. Typical symptoms include severe pain, blood in the urine, nausea, vomiting, fever, and chills. Every year, about 1.5 million people seek medical care for kidney stone problems. Kidney stones affect around 11% of men and 9% of women. Individuals with conditions such as high blood pressure, diabetes, and obesity are more susceptible to developing kidney stones. The objective of the proposed study is to detect the presence of kidney stones in Computed Tomography (CT) images of kidneys. The dataset is collected from the Kaggle website which contains 12446 CT images of kidneys. The film artefact is removed in the pre-processing stage. The pre-processing is divided into two steps: The median filter is used initially, and erosion is performed using structuring elements of three different sizes (3 x 3, 4 x 4, and 5 x 5). Better results are obtained for the 3x3 structuring elements. For image segmentation, the gradient vector flow method and the soft organ and bony skeleton removal approach are used. The features that are extracted from the segmented data include mean, standard deviation, contrast, correlation, energy, homogeneity, skewness, kurtosis, and entropy. The Butterfly Optimization Algorithm (BOA) is used for feature selection, while classification is done with the XG Boost supervised machine learning method