Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165

Kidney Stone Detection Using Image Processing and

Convolutional Neural Networks
PAVITHRA S SUPRAJA A
Department of Electronics and Communication Engineering, Department of Electronics and Communication Engineering,
Sri Sairam Engineering College, Sri Sairam Engineering College,
Chennai, India. Chennai, India.

SHANMUGAPRIYA G S. SARANYA
Department of Electronics and Communication Engineering, Department of Electronics and Communication Engineering,
Sri Sairam Engineering College, Sri Sairam Engineering College,
Chennai, India. Chennai, India.

Abstract:- Renal calculi, often known as kidney stones, rib pain, and abdominal pain. In addition to urinary tract
are solid masses made mostly of crystals. Detecting the infections, severe pain can also cause odorous nausea and
perfect and correct site of urinary calculus is essential for hematuria. The intensity of the pain varies and there are
surgical procedures. Because CT pictures include waves. Excessive urination, pollakiuria, and chills are all
greater speckle noise, manually detecting urinary calculi symptoms and symptoms of caution.
is difficult, hence automated systems for detecting kidney
stones in CT images are necessary. CT imaging is one of II. LITERATURE SURVEY
the imaging modalities available for diagnosing kidney
abnormalities, which can include changes in shape and According to the method proposed in [7], the
position, as well as swelling of the limb; other kidney superiority of self-stated KS records become decided the
abnormalities include the creation of stones, cysts, urine usage of NHANES weights and layout variables. An
obstruction, congenital defects, and malignant cells. This important position in this study is the increasing
research presents a new method for detecting kidney predominance of kidney stones in adult males as well as
stones. This project is divided into two parts: kidney CT females. Women of childbearing age have achieved the
image classification and kidney stone detection. For renal highest growth.
CT image classification, VGG16 convolutional neural
networks are employed, and for kidney stone Deep learning technologies and applications are the
identification, Fuzzy c means clustering is used. Filtering subject of the papers in the [17] special area. Deep learning is
and processing of kidney CT scans removes undesired a type of neural network with more layers allowing for greater
noise and enhances the image. Using complicated abstraction and better data predictions.
techniques such as the Convolutional Neural Network
(CNN) model, classify the image using the SoftMax Dayanand Jamkhandikar and his collaborators This
classifier algorithm, and ultimately detect the kidney approach converts images to grayscale and extracts features.
stone using FCM. This technology will produce more Brute pressure matchers test the traits of the database and go
accurate findings and will do it faster than the previous back the results.
way.
According to[1] the study was based on a cross-
Keywords:- Kidney Stones, Image Processing, CT image, sectional examination of NHANES responses from 2007 to
CNN. 2014. Data from four NHANES cycles (2007–2014) were
used to examine KS prevalence patterns in adults of various
I. INTRODUCTION ages. For the analysis, SAS version 9.4 was employed.

Kidney stones are tough composite of salts and III. SYSTEM ANALYSIS AND DESIGN
minerals along with calcium and urine that shape internal
your kidney. Body fluid content material is chargeable for Segmentation is essential in medical image processing.
the formation of those kidney stones. It takes time for our Traditional segmentation methods, such as thresholding and
body to become dehydrated and waste products to edge-based segmentation, are still in use, however they do not
accumulate. When the nutrient minerals in the urine become produce adequate segmentation results for medical images
supersaturated, stone crystals form. Kidney stones usually (Kidney Stone CT images).
spread with urine through the urethra to the bladder, but can
also enter the bladder and come out large. Stones are
regularly found by CT scans, x-rays, and ultrasound. The
most common symptoms of kidney stones are low back pain,

IJISRT22JUN1154 www.ijisrt.com 320

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
Machine gaining knowledge of strategies aren't as C. Classification:
correct as deep gaining knowledge of strategies.  After pretreatment of the transfer learning model Vgg16,
Nevertheless, gadget gaining knowledge of strategies are CNN is used to extract features and classify CT images of
used for category purposes. the kidney.
 The VGG16 network learns all of the info from a
IV. PROPOSED SYSTEM massive quantity of images. VGG16 is a 16-layer
convolutional neural network. The image provided as
The proposed system has following steps: input to the VGG network must be 224x224 pixels.
 The CT images for the kidney are taken for the public  After defining the network structure, you need to select
database. alternative training. For training purposes, the length
 Image analysis relies heavily on preprocessing. ought to be among 5 and 500. Network training makes use
Removing undesirable factors and additives from the of layer based architecture, schooling data, and schooling
renal CT image is the pre-processing stage. Therefore, parameters to classify the renal CT images.
you can make better decisions. To do this, we advise a  Finally, using the proposed vgg16 CNN model, test CT
method in which we use a median filter to clear noise and kidney images are categorized into two types, normal and
use adaptive histogram equalization to enhance picture pathological, based on their very rich function.
quality.
 After preprocessing, the VGG16 CNN network is utilized D. Kidney Stone Detection:
in getting to know a step to classify kidney CT images  Kidney stones are discovered following the
into regular and abnormal classes. categorization stage. In the kidney CT image, the FCM
 Finally, to detect kidney stones in abnormal kidney CT and level set segmentation techniques are employed to
images, we employed fuzzy c-means clustering(FCM) find the stone location.
and level set segmentation.  The technique of dividing the accumulated records or
gadgets into special subsets is known as as clustering
V. IMPLEMENTATION analysis
 The partition should have two properties:
 Modules Cluster homogeneity: To achieve cluster homogeneity,
Kidney stone detection done by following modules, data from the same cluster must be as identical as feasible.
A. Image Acquisition Data heterogeneity: To achieve data heterogeneity, data
B. Preprocessing from different clusters must be as different as possible.
C. Classification  The fuzzy c-means method (fuzzy c-means) is a
D. Kidney Stone Detection classification strategy that is based on the fuzzy
optimization of an aquadratic classification criterion. The
 Module Description: approach requires previous knowledge of the number of
A. Image Acquisition: classes and creates them iteratively using objective
 Image acquisition is the process of collecting an image function minimization.
from a source, often a hardware-based source, so that it  As a result, each pixel is allocated a degree of zone
may be processed later using whatever techniques are membership, resulting in a fuzzy image division. Finally,
required. in a clustered picture, the level set is employed to
 In image processing, image capture is always the first step recognise just the stone region
in the workflow, as processing is impossible without
images. The resulting image is not completely processed VI. SYSTEM ARCHITECTURE
and is the result regardless of the equipment used to make
it; this can be vital in some industries where having a
consistent foundation to work from is critical.
 The CT images for the kidney are taken for the public
database.

B. Preprocessing:
 The second stage of kidney stone detection is
pretreatment, including noise reduction, contrast
enhancement, and image scaling.
 The noise reduction part of preprocessing is a stage where
the noise from the renal CT image is done using the
median filter.
 Rectification is used to enhance the contrast of kidney CT
image via way of means of the use of Adaptive Histogram
Equalization (AHE).

Fig 1:- System Architecture

IJISRT22JUN1154 www.ijisrt.com 321

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
VII. OUTPUT

A. Normal Condition:

Fig 2.4:- Classification Using VGG16 CNN

Fig 2.1:- Input CT image

Fig 2.5:- Output for a Normal Kidney

B. Abnormal Condition:

Fig 2.2:- Noise Removal Using Median Filter

Fig 3.1:- Classification of an abnormal kidney

Fig 3.2:- Output for an abnormal kidney

Fig 2.3:- Contrast Enhanced Using AHE

IJISRT22JUN1154 www.ijisrt.com 322

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165

Fig 3.3:- Clustering of the Abnormal Kidney Image

Fig 3.6:- Detected Kidney Stone Image

VIII. CONCLUSION

This research describes two innovative algorithms for

categorising renal CT images and diagnosing kidney stones:
the VGG16 CNN transfer learning model and Fuzzy C-
means clustering. The primary purpose is to utilise Vgg16
CNN to classify normal and abnormal kidney CT images, as
well as to detect kidney stones in abnormal kidney CT
images using fuzzy c-means clustering and level set
segmentation. Modules for image capture, preprocessing,
classification, and kidney stone diagnosis are included in the
study. The suggested approach may distinguish between two
types of kidney CT pictures (normal/abnormal). Finally, the
experimental findings showed that the unique method beat
traditional tactics significantly.
Fig 3.4:- Clustered CT Image using Fuzzy C Means
REFERENCES

[1]. Chen Z, Prosperi M, Bird VY. Prevalence of kidney

stones in the USA: the National Health and Nutrition
Evaluation Survey. J Clin Urol 2018 Nov 26
[2]. G. Foster, C. Stocks, and M. Borofsky. Statistical Brief
#139: Emergency Department Visits and Hospital
Admissions for Kidney Stone Disease in 2009.
Statistical Briefs from the Healthcare Cost and
Utilization Project (HCUP). Agency for Healthcare
Research and Quality, Rockville, MD, 2012.
[3]. The American College of Radiology Criteria for ACR
appropriateness Acute flank pain with the possibility of
stone disease (urolithiasis).
[4]. AC Westphalen, RY Hsia, JH Maselli, R Wang, and R
Gonzales. National patterns, diagnoses, and predictors
of radiological imaging in patients with suspected
urinary tract stones. 2011;
[5]. Fwu CW, Eggers PW, Kimmel PL, Kusek JW, Kirkali
Z. Urolithiasis-related emergency department visits,
Fig 3.5:- Level Set Segmentation imaging, and medication usage have all increased in the
United States. Kidney International 2013

IJISRT22JUN1154 www.ijisrt.com 323

Volume 7, Issue 6, June – 2022 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
[6]. Wang DC, Parry CR, Feldman M, Tomlinson G,
Sarrazin J, Glanc P. Acute abdomen in the emergency
department: is CT a time-limiting factor? AJR Am J
Roentgenol 2015
[7]. Lee H, Tajmir S, Lee J, et al. Fully automated deep
learning system for bone age assessment. J Digit
Imaging 2017
[8]. Prevedello LM, Erdal BS, Ryu JL, et al. Automated
critical test findings identification and online
notification system using artificial intelligence in
imaging. Radiology 2017
[9]. Lakhani P, Sundaram B. Deep learning at chest
radiography: automated clas- sification of pulmonary
tuberculosis by using convolutional neural networks.
Radiology 2017
[10]. Levin S, Toerper M, Hamrock E, et al. Machine-
learning-based electronic triage more accurately
differentiates patients with respect to clinical out-
comes compared with the emergency severity index.
Ann Emerg Med 2018

IJISRT22JUN1154 www.ijisrt.com 324