Detection of Kidney Stone using Neural Network Classifier

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 06 | June 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2415
Detection of Kidney Stone using Neural Network Classifier
R. Devi1, Esther lipni.K2
1Assistant Professor, Computer science and Engineering, Sree Sastha Institute of Engineering and Technology,
Chennai India,
2 ME Final year Student, Computer science and Engineering, Sree Sastha Institute of Engineering and Technology,
Chennai India.
------------------------------------------------------------------------***------------------------------------------------------------------
Abstract— Kidney stone sickness is a typical issue among the western populace. Most kidney stones are little and pass
precipitously. These patients frequently need no further treatment. In any case, some nephrolithiasis patients foster enormous
stones, which can cause huge horribleness as intense side effects and ongoing confusions on the off chance that they are not
treated. However compelling treatment and counteraction might annihilate the sickness totally to defeat this we proposed
wavelet approach dodges both log and remarkable change, taking into account the completely evolved dot as added substance
signal-subordinate commotion with zero mean. The proposed strategy all through the wavelet change has the ability to
consolidate the data at various recurrence groups and precisely measure the neighborhood consistency of picture elements
and watershed calculation improve the picture in the quality manner and it characterizes with the Neural organization. Along
these lines, we really want to read up about picture handling for it. After that u will actually want to track down the cycle to
distinguish the kidney stone identification. For that reason, we are taking our popular philosophies for it. These are the initial
steps for our venture. Presently make a superior move to taking of info pictures utilizing CT pictures of the kidney from the
realized informational collection. Because of these means include for the recognition of the illness with the phases of the
sickness. Essentially, in kidney stone discovery will be finished by undesirable waste food like in the event that you are taking
tomato for day to day it will be impacted on kidney. To stay away from it by involving location in beginning phase utilizing
pre-process, division, highlight extraction of GLCM and brain network grouping of calculation.
INTRODUCTION
Kidney stone is a solid piece of material formed due to minerals in urine. These stones are formed by combination of
genetic and environmental factors. It is also caused due to overweight, certain foods, some medication and not drinking
enough of water. Kidney stone affects racial, cultural and geographical group.
Many methods are used for diagnosing this kidney stone such as blood test, urine test, scanning. Scanning also
differs in CT scan, Ultrasound scan and Doppler scan. Now days a field of automation came into existence which also being
used in medical field. Rather many common problems rose due to automatic diagnosis such as use of accurate and correct
result and also use of proper algorithms. Medical diagnosis process is complex and fuzzy by nature. Among all methods
soft computing method called as neural network proves advantages as it will diagnosis the disease by first learning and
then detecting on partial basis In this paper two neural network algorithms i.e. Feature extraction and watershed are used
for detecting a kidney stone. Firstly, two algorithms are used for training the data.
The data in the form of blood reports of various persons having kidney stone is obtained for various hospitals,
laboratories.
Fig .1. Structure of human body

Kidney stone is a solid piece of material formed due to minerals in urine. These stones are formed by combination of
genetic and environmental factors. It is also caused due to overweight, certain foods, some medication and not drinking
enough of water. Kidney stone affects racial, cultural and geographical group. Many methods are used for diagnosing this
kidney stone such as blood test, urine test, scanning.
Scanning also differs in CT scan, Ultrasound scan and Doppler scan. Now days a field of automation came into existence
which also being used in medical field. Rather many common problems rose due to automatic diagnosis such as use of
accurate and correct result and also use of proper algorithms. Medical diagnosis process is complex and fuzzy by nature.
Among all methods soft computing method called as neural network proves advantages as it will diagnosis the disease by
first learning and then detecting on partial basis In this paper two neural network algorithms i.e. Feature extraction and
watershed are used for detecting a kidney stone. Firstly, two algorithms are used for training the data. The data in the form
of blood reports of various persons having kidney stone is obtained for various hospitals, labo
Fig 2:
(a) Normal kidney (b) Damaged kidney
EFFECT OF KIDNEY STONE
Kidney can be effected by the chocolate, spinach, rhubarb, tea, and most nuts are rich in oxalate, poor diet, less
exercise, decreased sleep quality, an increase in caffeine intake, and unhealthy behavior also effected. Stones that are 4–6
mm are more likely to require some sort of treatment, but around 60 percent pass naturally.
This takes a normal of 45 days. So, we want know whether kidney is affected or not. For that process take kidney
detection. If it is affected then calculate the percent of effected area and what is the stage of the effected area. for that
process take neural network then we can detect the similar outputs.
SCOPE OF THE PROJECT
The main scope of the project is to predict the kidney stone from the person who is affected. And with help of
convolutional neural network technique.
In this we used discrete wavelet transform to identify the kidney stone.
Fig.3. Disease detection

EXISTING SYSTEM:
In existing system the threshold segmentation, cosine transformation and there used SVM algorithm and k means
clustering method. In existing system, the major problems are actual prosthesis are mostly limited to intrinsic visual and
acoustic feedback, available by observation of the prosthesis and sounds of the motors ,as it is the case for vibrotactors or
pressure. The most concerning issues is that most
Models have been tested extensively in controlled environments but the prove for robustness under the non-stationary
conditions of daily life is often missing. This needs to be reflected in the evaluation procedures. Furthermore, to achieve
real-time usability, appropriate design of the prosthetic device,
feature extraction and classification techniques should be properly investigated and implemented. The process should not
be in manual analysis, Continuous wavelet analysis and the distance of the wavelets is based on signal classification.
DISADVANTAGES OF EXISTING SYSTEM
 Accurate result was difficult to get
 In a small time multiple images are cannot be detected
 Inaccuracies classification are done because the image contains a noise by operator performance in medical
resonance images.
Fig.4. Abnormalities Detection image
PROPOSED SYSTEM:
In proposed system we using DWT, preprocessing images and gray level co-occurance martic is used for accuracy.
In proposed system, the problems in existence are overcome by, some techniques . It becomes a pre processing system
and the system has advanced to use Discrete wavelet analysis instead of Continuous wavelet analysis after investigation.
Gray scale Co- occurrence matrix is introduced so , the accuracy of the signals or wavelet can be efficient. The sound of the
machines is reduced because of using EMG ports. They can be used to give a accurate wavelet signal in a discrete form.
ADVANTAGES OF PROPOSED SYSTEM
 It is easily identify the kidney stone using neural network and high accuracy rate is proposed in this project.
 The algorithm Proves to be simple andeffective
 Gray scale Co-occurrence matrix extracts the features accurately.

Fig.5 proposed system diagram
RELATED SEARCH
DEEP LEARNING
Profound learning (otherwise called profound organized learning) is important for a more extensive group of AI strategies
in view of counterfeit brain networks with portrayal learning.
Learning can be managed, semi-directed or unaided. Profound learning structures, for example, profound brain
organizations, profound conviction organizations, repetitive brain organizations and convolutional brain networks have
been applied to fields including PC vision, discourse acknowledgment, normal language handling, sound acknowledgment,
informal organization separating, machine interpretation, bioinformatics, drug plan, clinical picture examination, material
assessment and prepackaged game projects, where they have delivered results practically identical to and at times
astounding human master execution. The adjective "deep" in deep learning comes from the use of multiple layers in the
network. Early work showed that a linear perceptron cannot be a universal classifier, and then that a network with a
nonpolynomial activation function with one hidden layer of unbounded width.
Fig 6 : CNN classifier

EXPERIMENT PLATFORM
The first step in establishing a predictive model for decoding multi-DOF activities (involving either motion or force) is to
collect a sufficient number of training samples. Thus, supervised wrist activities are detected simultaneously along with
the raw EMG signals. In previous work, we developed a platform for collecting these signals contains a wrist force-to-
movement mapping device (FMM) of our own design.
The springs on the FMM make it possible to map the force and position of the wrist joint. The subjects must exert
sufficient force to perform various wrist movements in a semi-constraint manner.
In this way, the movements mimic closely the natural situation during reach-and-grasp activities, and the elicited multi-
DOF wrist force can be quantified accurately through the relevant ones. A laser on the device makes it possible to interpret
the extension-flexion (E-F), ulnar-radial deviation (U-R), and supination-pronation (S-P) movements of the wrist in terms
of the horizontal (x-axis), vertical (y-axis), and rotational (z- axis) motions of a cross-shaped cursor projected onto a
screen.
DIGITAL IMAGE PROCESSING
The identification of objects in an image and this process would probably start with image processing techniques
such as noise removal, followed by (low-level) feature extraction to locate lines, regions and possibly areas with certain
textures.
IMAGE
A picture is a two-layered picture, which has a comparable appearance to some subject normally an
actual item or an individual.
IMAGE TYPES
RGB: - these are red, green and blue. A (computerized) variety picture is an advanced picture that incorporates
variety data for every pixel. Every pixel has a specific worth which decides its seeming tone.
This worth is qualified by three numbers giving the decay of the variety in the three essential tones Red, Green
and Blue. Any color visible to human eye can be represented this way. The decomposition of a color in the three primary
colors is quantified by a number between 0 and 255. For example, white will be coded as R = 255, G = 255, B = 255; black
will be known as (R, G, B) = (0,0,0); and say, bright pink will be: (255,0,255).
MODULE 1: PROCESSING INPUT
Fig 6 CT examined pictures
DESCRIPTION:
Bringing in the picture through picture securing apparatuses; Breaking down and controlling the picture; Output in
which result can be adjusted picture.

Image Pre-processing is a common name for operations with images at the lowest level of abstraction. Its input and
output are intensity images.
The point of pre-handling is an improvement of the picture information that stifles undesirable contortions or upgrades
some picture highlights significant for further processing.
MODULE 2: BPNN
Fig 7 BPNN
DESCRIPTION:
Back-propagation is the essence of neural net training. It is the method of fine-tuning the weights of a neural net
based on the error rate obtained in the previous epoch (i.e., iteration). Proper tuning of the weights allows you to reduce
error rates and to make the model reliable by increasing its generalization. Backpropagation is a short form for "backward
propagation of errors." It is a standard method of training artificial neural networks.
This method helps to calculate the gradient of a loss function with respects to all the weights in the network. Simplifies
the network structure by elements weighted links that have the least effect on the trained network.
You need to study a group of input and activation values to develop the relationship between the input and hidden unit
layers. It helps to assess the impact that a given input variable has on a network output.
The knowledge gained from this analysis should be represented in rules. Backpropagation is particularly helpful for
profound brain networks chipping away at mistake inclined projects, like picture or discourse acknowledgment.
Backpropagation exploits the chain and power rules permits backpropagation to work with quite a few results.
MODULE 3: GLCM FEATUES
Fig 8: Bit grayscale image & gray levels

DESCRIPTION:
To make a GLCM, utilize the graycomatrix capability. The graycomatrix capability makes a dim level co-event framework
(GLCM) by working out how frequently a pixel with the force (dim level) esteem I happens in a particular spatial
relationship to a pixel with the worth j.
Of course, the spatial relationship is characterized as the pixel of interest and the pixel to its nearby correct (on a level
plane contiguous), yet you can determine other spatial connections between the two pixels.
Every component (i,j) in the resultant GLCM is essentially the amount of the times that the pixel with esteem I happened in
the predefined spatial relationship to a pixel with esteem j in the information picture. Since the handling expected to work
out a GLCM for the full powerful scope of a picture is restrictive, graycomatrix scales.
the input image. By default, graycomatrix uses scaling to reduce the number of intensity values in gray scale image from
256 to eight. The number of gray levels determines the size of the GLCM.
To control the number of gray levels in the GLCM and the scaling of intensity values, using the Num Levels and the Gray
Limits parameters of the graycomatrix function. See the graycomatrix reference page for more information.
MODULE 4: DWT
Fig 9: DWT
DESCRIPTION:
The DWT and the discrete wavelet transforms differ in how they discretize the scale parameter. The CWT
typically uses exponential scales with a base smaller than 2, for example 21/12 . The discrete wavelet transform always
uses exponential scales with the base equal to 2.
The scales in the discrete wavelet transform are powers of 2. Keep in mind that the physical intrepretation of scales
for both the CWT and discrete wavelet transforms requires the inclusion of the signal’s sampling interval if it is not equal
to one. For example, assume you are using the CWT and you set your base to s0=21/12.
To attach physical significance to that scale, you must multiply by the sampling interval Δt, so a scale vector covering
approximately four octaves with the sampling interval taken into account is sj0Δt j=1,2,⋯48. Note that the sampling
interval multiplies the scales, it is not in the exponent. For discrete wavelet transforms the base scale is always 2.
The decimated and nondecimated discrete wavelet transforms differ in how they discretize the translation parameter.
The decimated discrete wavelet transform (DWT), always translates by an integer multiple of the scale, 2jm . The
nondecimated discrete wavelet transform translates by integer shifts.
The scales in the discrete wavelet transform are powers of 2. Keep in mind that the physical intrepretation of
scales for both the CWT and discrete wavelet transforms requires the inclusion of the signal’s sampling interval if it is not
equal to one. For example, assume you are using the CWT and you set your base to s0=21/12.

MODULE 5 : NEURAL NETWORK
Fig 10 : NN
DESCRIPTION:
The assurance of the name "cerebrum association" was one of the exceptional PR accomplishments of the Twentieth
Century. It irrefutably sounds more stimulating than a specific depiction, for instance, "An association of weighted, added
substance values with nonlinear trade capacities". Regardless, no matter what the name, cerebrum networks are quite far
from "thinking machines" or "phony personalities". An ordinary phony mind association could have 100 neurons. In
relationship, the human tactile framework is acknowledged to have around 3x1010 neurons. We are still light seemingly
forever from "Data".
The first "Perceptron" model was created by Frank Rosenblatt in 1958. Rosenblatt's model comprised of three layers, (1) a
"retina" that circulated contributions to the subsequent layer, (2) "affiliation units" that join the contributions with loads
and trigger a limit step capability which feeds to the result
layer, (3) the result layer which joins the qualities. Tragically, the utilization of a stage capability in the neuronsmade the
discernments troublesome or difficult to prepare. A basic examination of perceptrons distributed in 1969 by Marvin
Minsky and Seymore Paper brought up various basic shortcomings of perceptrons, and, for a while, interest in perceptrons
disappeared.
CONCLUSION:
The conclusion is drawn as the solution we got from the previous project to now the changes we made it shows the
difference in accuracy and the number of images can be checked in a single timing.
And we used BPNN instead of SVM. In future we were tried to increased the performance of this process and able to get
more accuracy. The proposed work is advantageous for recognizing kidney stones from CT scan pictures with less
processing instant and achieves great accuracy.
REFERENCES:
[1] W. -Y. Chung, R. F. Ramezani, C. -H. Li, V. F. Tsai and M. Mayeni, "Development of Low-voltage Urine Sample
Conductivity Measurement System for Kidney Stone Risk Assessment," 2021 IEEE 3rd Eurasia Conference on Biomedical
Engineering, Healthcare and Sustainability (ECBIOS), 2021, pp. 13-15, doi: 10.1109/ECBIOS51820.2021.9510254.
[2] A. Soni and A. Rai, "Kidney Stone Recognition and Extraction using Directional Emboss & SVM from Computed
Tomography Images," 2020 Third International Conference on Multimedia Processing, Communication & Information
Technology (MPCIT), 2020, pp. 57-62, doi: 10.1109/MPCIT51588.2020.9350388.
[3] N. Thein, H. A. Nugroho, T. B. Adji and K. Hamamoto, "An image preprocessing method for kidney stone segmentation in
CT scan images," 2018 International Conference on Computer Engineering, Network and Intelligent Multimedia (CENIM),
2018, pp. 147-150, doi: 10.1109/CENIM.2018.8710933.

2018, pp. 147-150, doi: 10.1109/CENIM.2018.8710933.
2018, pp. 147-150, doi: 10.1109/CENIM.2018.8710933.

Detection of Kidney Stone using Neural Network Classifier

More Related Content

Similar to Detection of Kidney Stone using Neural Network Classifier (20)

More from IRJET Journal (20)

Recently uploaded (20)

Detection of Kidney Stone using Neural Network Classifier