A SURVEY ON KIDNEY STONE DETECTION USING IMAGE PROCESSING AND DEEP LEARNING
0 likes917 views
The document describes a study on detecting kidney stones using image processing and deep learning techniques. The researchers preprocessed CT and MRI scan images of kidney stones using techniques like gray level co-occurrence matrix for feature extraction. They then trained a convolutional neural network (CNN) model on the images to classify kidney stones. The CNN model consisted of convolution, ReLU, pooling and fully connected layers. The trained model could accurately detect and classify kidney stones in input images. The researchers concluded that combining image processing and deep learning was an effective method for automated kidney stone detection.
![International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1053
A SURVEY ON KIDNEY STONE DETECTION USING IMAGE
PROCESSING AND DEEP LEARNING
Apoorva Mohite1, Aishwarya Avatade2, Meghana Galande3, Prof. Ganesh V. Madhikar4
1-3Student, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering,
Maharashtra, India.
4Assistant Professor, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering,
Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract – As result of our current life style kidney stone
has become a common health issue. There are inaccuracies in
the classification of kidney stone due to the presence of noise.
Also, quick and correct diagnosis of kidney stone is essential
which is observed to be lacking in the currently followed
practices.
It's difficult to obtain results for large dataset using human
inspection, this is where an automated kidney stone
classification is implemented. The automated system uses
image processing and deep learning method. The MIR and CT
scan images of the proposed methodology of nephrolithiasis is
preprocessed. The extraction of the key features is done using
gray level concurrent matrix. The conversion of RGB format of
image into gray format is essential. The color information of
the image is now reduced and converted into a single
dimensional from 3 dimensional with similar patterns.
Key Words: Image processing, convolution neural
network, Deep learning, Machine learning, Python, CT
scans.
1. INTRODUCTION
The Kidney Stone issue can be seeing rising dramatically
throughout the world. Shapeofkidneysarelikebeanshaped.
They are located on both side of spine behind bellies and
below the ribs. Size of kidney is around size of a largest fist.
Filtration of the blood is the primary function of kidneys.
They maintain balance of bodily fluids by removing waste
materials from it. Also, they keep electrolytes in their
sufficient levels. When blood comes into kidney the work of
kidney starts like removing waste and adjusting level ofsalt,
water and minerals if it is needed. Then this filtered blood
goes into body back and the waste goes into pelvis and then
removed from body in the form of urine funnel shaped
structure that drains down a tube known as ureter to the
bladder.
Each and every kidney stone having around ten percent tiny
filters. They are known as nephrons. If blood stops flowing
through kidney part it could be die and that can lead to a
kidney failure. Formation of a stone in kidney leads to
blockage of urine congenital anomalies cysts. Various types
of kidney stones namely viz renal calculi stone, struvite
stones, stage horn was analysed. A commendable
contribution of various researcher in the discipline of
nephrolithiasis detection via means of occurring numerous
algorithms to locate the kidney stone is seen. Use of neural
network for the classification of urinary calculus has shown
great potential.
1.1 Problem Statement
Failure of Kidney can be a life changing. So that the initial
detection of kidney stone is important. Kidney stones must
first be identified to ensuresuccessfulsurgicaloperations.[3]
1.2 Objective
The main objective of this project is to efficiently detect
kidney stone problems with the help of image, and to
improve the detection rate in terms of accuracy as well as
sensitivity.
2. LITERATURE SURVEY
For this topic, many research papers have been published
and many researchers have work upon it, in order to design
Kidney stone detection using image processing and deep
learning few of the following are discussed here. Literature
survey is an information review. Which will help us in
understanding and exploring concept of basis learningsSo it
will help us in better understanding the topics based on
early information available. Literature survey is often done
to connect our work with the relation of existing data.](https://image.slidesharecdn.com/irjet-v9i5273-220930110202-198faf6a/85/A-SURVEY-ON-KIDNEY-STONE-DETECTION-USING-IMAGE-PROCESSING-AND-DEEP-LEARNING-1-320.jpg)
A SURVEY ON KIDNEY STONE DETECTION USING IMAGE PROCESSING AND DEEP LEARNING
- 1. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1053 A SURVEY ON KIDNEY STONE DETECTION USING IMAGE PROCESSING AND DEEP LEARNING Apoorva Mohite1, Aishwarya Avatade2, Meghana Galande3, Prof. Ganesh V. Madhikar4 1-3Student, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India. 4Assistant Professor, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract – As result of our current life style kidney stone has become a common health issue. There are inaccuracies in the classification of kidney stone due to the presence of noise. Also, quick and correct diagnosis of kidney stone is essential which is observed to be lacking in the currently followed practices. It's difficult to obtain results for large dataset using human inspection, this is where an automated kidney stone classification is implemented. The automated system uses image processing and deep learning method. The MIR and CT scan images of the proposed methodology of nephrolithiasis is preprocessed. The extraction of the key features is done using gray level concurrent matrix. The conversion of RGB format of image into gray format is essential. The color information of the image is now reduced and converted into a single dimensional from 3 dimensional with similar patterns. Key Words: Image processing, convolution neural network, Deep learning, Machine learning, Python, CT scans. 1. INTRODUCTION The Kidney Stone issue can be seeing rising dramatically throughout the world. Shapeofkidneysarelikebeanshaped. They are located on both side of spine behind bellies and below the ribs. Size of kidney is around size of a largest fist. Filtration of the blood is the primary function of kidneys. They maintain balance of bodily fluids by removing waste materials from it. Also, they keep electrolytes in their sufficient levels. When blood comes into kidney the work of kidney starts like removing waste and adjusting level ofsalt, water and minerals if it is needed. Then this filtered blood goes into body back and the waste goes into pelvis and then removed from body in the form of urine funnel shaped structure that drains down a tube known as ureter to the bladder. Each and every kidney stone having around ten percent tiny filters. They are known as nephrons. If blood stops flowing through kidney part it could be die and that can lead to a kidney failure. Formation of a stone in kidney leads to blockage of urine congenital anomalies cysts. Various types of kidney stones namely viz renal calculi stone, struvite stones, stage horn was analysed. A commendable contribution of various researcher in the discipline of nephrolithiasis detection via means of occurring numerous algorithms to locate the kidney stone is seen. Use of neural network for the classification of urinary calculus has shown great potential. 1.1 Problem Statement Failure of Kidney can be a life changing. So that the initial detection of kidney stone is important. Kidney stones must first be identified to ensuresuccessfulsurgicaloperations.[3] 1.2 Objective The main objective of this project is to efficiently detect kidney stone problems with the help of image, and to improve the detection rate in terms of accuracy as well as sensitivity. 2. LITERATURE SURVEY For this topic, many research papers have been published and many researchers have work upon it, in order to design Kidney stone detection using image processing and deep learning few of the following are discussed here. Literature survey is an information review. Which will help us in understanding and exploring concept of basis learningsSo it will help us in better understanding the topics based on early information available. Literature survey is often done to connect our work with the relation of existing data.
- 2. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1054 c l 3. SOFTWARE REQUIRMENTS 3.1 PYTHON In this project we use Python language for coding. It is High level and free open-source language. Specifications: - High level Interpreted Easy to Code Interpreted and Portable language GUI programming support Inbuild Library support 3.2 VS CODE IDE: - VS code also known as Visual Studio Code it is an IDE made by Microsoft used for development operations like task running, debugging. It’s aim to provide tools that developer needs for a quick code-build-debugcyclethatdeveloperscan write and test code at the same time. Specifications: - Lightning-fast source code editor Syntax highlighting Auto indentation Snippets 3.3 CNN ALGORITHM: - A CNN algorithm also known as Convolutional Neural Network is a Deep learning algorithm. That take the input images and assign importance to various aspects so be able to differentiate one from another.ACNN hasmultiplehidden layers that help to extract information from an image. The four layers of CNN are: 1. Convolution layer 2. Relu layer 3. Pooling layer 4. Fully connected layer 3.4 Machine Learning: - Machine learning allow user to feed an algorithm with an large amount of data and have the computer analyze and make data-driven recommendations and decisionsbased on only the input data, for example -An algorithm would be trained with pictures of dogs and other things,all labelledby humans, and the machine would learn ways to identify pictures of dogs on its own. Types of machine learning algorithms: • Supervised • Unsupervised Supervised machine learning is the most common and easy type. Machine learning algorithms are utilized during a good kind of applications, like in medicine, email filtering, speech recognition, and computer vision, where it's difficult or unfeasible to develop conventional algorithms to perform the needed tasks. 3.5 Image Processing: - It is may be a method to convert a picture into digital form and perform some operations thereon, so as to urge an enhanced image or to extract some useful information from it. Morphological image processing removes the imperfections from the binary images because binary regions produced by simple thresholding it can be distorted by the noise. It helps in smoothing the image using opening as well as closing operations. Morphological operations can be extended to grayscale images. It consists of non-linear operations associated with the structure of features of a picture. It depends on related ordering of pixels but on their numerical values. This technique analyzes an image using a small template known as structuring element and this element is placed on different possible locations in the image and is compared with the corresponding neighborhood pixels. A small matrix structuring element is with 0 and 1 values.
- 3. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1055 Fig 3.5 Image Processing 4. DESIGN AND IMPLEMENTATION 4.1 FLOW CHART: Fig.4.1 Flow chart Fig.4.1 shows flow chart of we can see the work flow of our project system. Firstly, CNN training is done with the help of Machine learning and Deep learning by doing processing in dataset. So, for that dataset of stone and no stone both images are passed and after passing the data arrangement on data is done and arrange dataset is passed for image processing. Here resizing images, setting flag to images and labelling images these operations are done.Theseprocessed images are then sent for training. Using CNN algorithm, we train the data. Here 3 layers of CNN are used and feature extraction of images is done. So, after doing these all operations these all is store inonetrainedmodule.Intrained module one input is from user input image is coming and this image is compared with training did on dataset. And according to comparison predicted result is generated whether stone is present or not and finally output is display. 4.2 BLOCK DIAGRAM: Fig.4.2. Block Diagram The following block diagram Fig.4.2 shows implementation of kidney stone detection. Which containsCNN algorithmfor training. The following diagram shows flow of signal. The main function of CNN algorithm istotakeaninputimage and to analyze visual images by processing data with grid-like topology. Image processing block is used to process images in given dataset like labeling images, resizing the images setting flag to images, feature extraction etc. like operations are done in this block. Coding is done in Python language. The main function of CNN training is to create neural network and creating datasetfortrainingandtesting.InCNN trained module comparison is done and result is predicted and accordingly result is display.
- 4. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1056 5. OUTPUT: 6. CONCLUSIONS Detecting the presence of kidney stones using the proposed methodologyhasbeendone bypreprocessingtheultrasound image. It was followed by segmentation and finally morphological analysis of the resulting image was performed. The final image helped in the detection of the exact location of the stone. Moving further the edge detection method was performed which identified the shape and structure of the formed stones. 7. FUTURE SCOPE • In future work, the proposed method might be designed for real time implementation via interfacing it with the scanning machines. • In future, the system will be designed for real time application by placing biomedical sensors in the abdomen region to capture kidney portion.Thecapturedkidneyimage is to be proposed algorithm to process and detect stone on FPGA using hardware description language (HDL). The identified urinary calculuswithintheimageisdisplayedwith color for straightforwardidentificationandvisibilityof stone in monitor. REFERENCES 1. Malathy Chidambaranathan and Gayathri Mani "kidney stone detection with CT images using neural network" Research gate May 2020 DOI:10.37200/IJPR/V24I8/PR280269Conference: iciot2020 2. T. Vineela, R. V. G. L. Akhila, T. Anusha, Y. Nandini, S. Bindu "Kidney Stone Analysis Using Digital Image Processing" International Journal of Research in Engineering, Science and Management Volume-3, Issue-3, March-2020 www.ijresm.com | ISSN (Online): 2581-5792 3. Kalannagari Viswanath and Ramalingam Gunasundari "Analysis and Implementation of Kidney Stone Detection by Reaction Diffusion Level Set Segmentation Using Xilinx System Generator on FPGA", Hindawi Publishing Corporation VLSI Design, Volume 2015, Article ID 581961, http://dx.doi.org/10.1155/2015/581961 4. Anushri Parakh, Hyun Kwang Lee, Jeong Hyun Lee, Brian H. Eisner, Dushyant V. Sahani and Synho Do "Urinary Stone Detection on CT Images Using Deep Convolutional Neural Networks: Evaluation of Model Performance and Generalization" Radiol Artif Intell. 2019 Jul; 1(4): e180066.Published online 2019 Jul 24. DOI: 10.1148/ryai.2019180066 5. Venkatasubramani.K, K. Chaitanya Nagu,P.Karthik, A. Lalith Vikas, “Kidney Stone Detection Using Image Processing and Neural Networks”, Annals of R.S.C.B., ISSN:1583-6258, Vol. 25, Issue 6, 2021 6. Mrs. Monica Jenifer J, A Roopa, C R Sarvasri, G Sharmila, A Yamuna "Design andimplementation of kidney stones detection using image processing technique" International Research Journal of Engineering and Technology (IRJET) e- ISSN: 2395-0056 Volume: 08 Issue: 05 | May 2021 www.irjet.net p-ISSN: 2395-0072
- 5. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 09 Issue: 05 | May 2022 www.irjet.net p-ISSN: 2395-0072 © 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1057 BIOGRAPHIES Apoorva Pramod Mohite Student, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India. Aishwarya Arun Avatade Student, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India. Prof. Ganesh V. Madhikar Assistant Professor, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India Meghana Deepak Galande Student, Dept. of Electronics and Telecommunication Engineering, Sinhgad College of Engineering, Maharashtra, India.