Segmentation of Renal Calculi in Ultrasound Kidney Images Using Modified Watershed Method

Introduction

The most widespread problem in the human urinary system is renal calculus which is also known as kidney stones or urinary stones (Manousakas, Lai, & Chang, 2010). Though considerable suffering and at times renal failure are caused by kidney stone diseases that occur in roughly 10% of the U.S. population affected, the mechanism for this disease is not adequately known (He, Deng, & Ouyang, 2010). The principal organ in the urinary system namely, kidney not only produces urine but it is also useful in purifying the blood. Disposing poisonous substances from the blood and maintaining the useful components in proper balance are the two essential functions of kidney. Among the diverse types of kidney stones that exist, calcium-containing stones, Uric acid stones, Struvite, or infected stones and Cystine stones are four fundamental types (Shah, Desai, & Panchal, 2010). Kidney diseases are commonly classified into hereditary, congenital or acquired (Bommanna, Madheswaran, & Thyagarajah, 2010). The identification of calcifications inside the body is a large field of study including many dynamic areas of research, which is especially useful for diagnosing the kidney stone diseases. Prominent effects that are utilized to detect fracture in real kidney stones that can have arbitrary non-spherical shape are related to the reverberation time across the length of the stone (Manousakas et al., 2006).

Strong speckle noise and attenuated artifacts present in the abdominal ultrasound images poses a unique challenge in using these images for stone segmentation (Gupta, Gosain, & Kaushal, 2010). Accordingly, this task entails the usage of as much beforehand knowledge as possible, like texture, shape, spatial location of organs and so on. The performance of the several automatic and semiautomatic methods that have been proposed deteriorates quickly when the structures are inadequately defined and have low contrast like the neuroanatomic structures, such as thalamus, globus pallidus, putamen, etc., though it has normally a good performance when the contrast-to-noise ratio is high (Maulik, 2009). Regular clinical practice Numerous extensively available medical imaging methods like X-ray, positron emission tomography (PET), computer tomography (CT), Ultrasound (US) and magnetic resonance imaging (MRI) are extensively used in regular clinical practice (Jouannot et al., 2004). As compared to other medical imaging modalities such as computed tomography (CT) and magnetic resonance imaging (MRI), the US is particularly difficult to segment because Ultra Sound is especially hard to segment because its image quality is somewhat low compared to other medical imaging modalities like computed tomography (CT) and magnetic resonance imaging (MRI) (Xie, Jiang, & Tsui, 2005). Quality of image data significantly affects the segmentation of Ultra Sound (US) images (Noble & Boukerroui, 2006). It is difficult to extract features that describe the kidney tissues by segmenting the kidney area (Bommanna, Madheswaran, & Thyagarajah, 2010). Despite this, ultrasound images are extensively used in the medical field (Jeyalakshmi & Kadarkarai, 2010). Tamilselvi (2013) proposed the segmentation method using Squared Euclidean Distance (SED) with ANFIS in supervised learning has made the technique more efficient than the previous techniques. Thus the obtained error is minimized when compared to the existing algorithm that leads to high efficiency.