IRJET- Lung Cancer Detection using Grey Level Co-Occurrence Matrix

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 04 | Apr 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 2191
Lung Cancer Detection Using Grey Level Co-Occurrence Matrix
Mahpara Alam1, Bhavika Chaudhari2, Shamal Bangar3, Pratiksha Desale4, Prof. Chaitali Patil5
1,2,3,4,Students, Department of Computer Engineering, K.K.W.I.E.E.R, Maharashtra, India
5Professor, Department of Computer Engineering, K.K.W.I.E.E.R, Maharashtra, India
---------------------------------------------------------------------***----------------------------------------------------------------------
Abstract - Lung cancer is curable disease if detected in its
early stages. In recent years, Image processing techniques are
widely used in medical domain for image improvement in
earlier detection. For diagnosis of lung cancer, CT image of
lung is one of the most common imaging modalities. Manually
identifying tumor from hundreds of CT images for any patient
may prove to be very tedious and time consuming task. Hence
this system aims at detection of cancer through an automated
process to minimize human error and making the process
more accurate and hassle-free. Lung carcinomas are
categorized by the size and appearance of the malignantcells.
The growth of epithelial cells on lung parenchyma region
shows in CT image which can be used in the system. In this
proposed approach image pre-processing techniquesareused
like noise removal, enhancement and image segmentation
then the feature extraction recognize the region of interest of
tumor. SVM (Support Vector Machine) algorithm can be used
for classification. GLCM (Grey Level co-occurrence Matrix)
based textual features detection can be usedtosearchthelung
parenchyma region and extract the features.
Key Words: GLCM, CT images, SVM, Feature extraction,
Image processing
1. INTRODUCTION
Manually identifying tumours from hundreds of MRI slices
for any patient may prove to be very tedious and time
consuming task. Cancer is curable disease if it detected in its
early stage. To reduce the death rates due to occurrence of
lung cancer we need some methodology to detect the cancer
in its early stage. Not only the detection of diseaseinitsearly
stage may reduce the death rate but also proper diet and
doctor treatment can be done on time.
Detecting Cancer is still challenging for the doctors in the
field of medical. Even now the actual reason and complete
cure of cancer is not invented. For better solution we are
going to design and implement lung cancerdetectionsystem
to minimize human error and makingprocessmoreaccurate
and hassle-free.
2. LITERATURE SURVEY
Mokhled S. AL-TARAWNEH [1], The study on the image
processing techniques is carried out for detection of cancer.
According to author many techniques are available todetect
cancer earlier, but none of the technique gave good result
because of time factor. Image quality and accuracy was the
core factor of this research. Imagequalityassessmentaswell
as improvement depends on the enhancement stage where
pre-processing techniques was used.
Ritika Agarwal, Ankit Shankhadhar, Raj Kumar Sagar [2],
Many techniques are available to determine the existence of
nodule in the lung in its early stage but noneofthetechnique
gave the best result. Therefore theyimplementedthesystem
to give more accuracy and best result they propose a system
called as Computer Aided diagnosis System CAD for
identifying the presence of nodule in lung.
Khinn Mya Mya Tun [3] , The implementation of image pre-
processing and image segmentation is carried out to obtain
the diagnosis result. Median filter was used for image pre-
processing. In feature extraction Gray Level Co-Occurrence
matrix (GLCM) was used. In classification, feed-forward
neural network was used to classify the lung cancer stages.
Mr. Akshay Bhor, Mr. Aditya Likhar, Mr. Azhar Maner, Mr.
Aditya Patil, Prof. Dipti Chaudhari [4], The main aim of the
system was to automate the classification method for the
first detection of carcinoma. It included classification
algorithmic program i.e. Neural Network and for
improvement GA (Genetic Algorithm) was employed.
Tanushree Sinha Roy, Neeraj Sirohi, Arti Patle[5], focuses
their study mainly on the classification of lung images as
affected by cancer disease or not. In their system lung CT
scan image was used as input image and then some
important features was extracted to classify the image. To
classify the image Fuzzy Interface System was used. The
system used both neural network and fuzzy logic technique.
3. PROPOSED SYSTEM
Fig -1: Proposed System

3.1 Image Database
The required database for system must contain the CT scan
Images. The CT scan image is more reliable than other form
of images. The images are in the DICOM format. It needs to
be converted to use as input to the system.
3.2 Image Preprocessing
Smoothing, image enhancement are the techniques areused
in pre-processing. In proposed model for median filter is
used to remove the noise from the given image.Medianfilter
give the best result because it removes the noise without
blurring the image whereas other techniques remove the
noise but give blur image as output. Image enhancement
improves the visibility of features for better understanding.
It also highlights the high frequency area and suppress the
low frequency area.
3.3 Feature Extraction
The features like energy, entropy,homogeneity,contrast can
be extracted after the segmentation process is done in the
region of lung, for analyzing some diagnosis to identify the
cancer in the lung correctly. In this study, the features are
used by representing Gray level Co-occurrence matrix.
4. GLCM
The extraction of the texture features using GLCM has been
considered as the significant technique and it has been used
in several applications of remote sensing for analysis of the
texture.
GLCM points to Gray level Co-occurrence matrix. It is of the
second order statistics, so informationwith regardstopixels
of pairs are collected by GLCMalgorithm.GLCMexhibitshow
the pixel brightness occurs in an image. A matrix is built up
at a distance of d=1 and at angles in degrees (0, 45, 90, 135).
GLCM texture picks up the relation between two pixels at a
time, called the neighbour and the reference pixel. GLCM
calculates the distance and angular spatial relationship over
an image sub- region of specific size. GLCM is prepared by
using gray scale values.
Fig -2: Test image figure
Fig -3: GLCM Calculation
The extraction of features is done as the follows:
1. Divide the planes of the image into RGB
2. Calculate the matrices of GLCM as per given equation.
3. Calculate the geometric features for the obtained GLCM
matrix:
Energy, Contrast, Entropy, Homogeneity.
Notations:
Pij = Element ij of GLCM
N = No. of Gray levels in the image
4.1 Energy
This statistic is also called as Angular second moment or
Uniformity. It measures the textural uniformity that is the
pixel pair repetitions. It detects disorders present in
textures. Energy reachesa maximumvalueequal tothevalue
one. High energy values occur when the distribution of gray
level has a constant or periodic form.
0 0 1 1
0 0 1 1
0 2 2 2
2 2 3 3
Neighbour pixel value-
Reference pixel value:
0 1 2 3
0 (0,0) (0,1) (0,2) (0,3)
1 (1,0) (1,1) (1,2) (1,3)
2 (2,0) (2,1) (2,2) (2,3)
3 (3,0) (3,1) (3,2) (3,3)

4.2 Entropy
This statistic measures the complexity of an image. The
entropy is large when the image is not texturally uniform
and many of the GLCM elements have very small values.
Complex textures tend to have a high entropy. Entropy is
strongly, but inversely correlated to the energy.
4.3 Contrast
This statistic measures the spatial frequency of an image
which is the difference moment of GLCM. It is the difference
between the highest and the lowest values of a contiguous
set of pixels present in an image.
4.4 Homogeneity
This statistic is also called as the InverseDifferenceMoment.
It measures image homogeneity. It is sensitive to the
presence of near diagonal elements in the GLCM. When all
elements in the image are same it gives maximum value.
GLCM homogeneity and contrast are strongly, but inversely
correlated in terms of equivalent distribution in the pixel
pair population. It means if contrast increases homogeneity
decreases while energy is kept constant.
5. CONCLUSIONS
Lung Cancer is one of the largest occurred disease in the
world. Only early detection of disease may help to decrease
the death rate due to the lung cancer. Image processing
techniques are widely used in some medical areas for image
improvement in earlier detection and treatment stages,
where the time factor is very important to discover the
abnormality issues in target images.
This proposed system is able to detect the cancer in early
stages more correctly. We proposed the methodology to
detect the cancerous lung CT image based on GLCM.
REFERENCES
[1] Mokhled S. AL-TARAWNEH,”Lung Cancer Detection
Using Image Processing Techniques”. Leonardo
Electronic Journal of Practices and Technologies,2012
[2] Ritika Agarwal, Ankit Shankhadhar, Raj Kumar Sagar
,”Detection of Lung Cancer using content based Medical
Image Retrieval ,Fifth International Conference on
Advanced Computing and Communication
Technologies,2015
[3] Khinn Mya Mya Tun ,Feature Extraction and
Classification of Lung Cancer Nodule using Image
ProcessingTechniques,International Journal ofAdvance
Research in Engineering, Science & Technology,2014
[4] Mr. Akshay Bhor, Mr. Aditya Likhar , Mr. Azhar Maner ,
Mr. Aditya Patil, Prof. Dipti Chaudhari , Classification of
lung cancer stages using Quantitative Image and
Genomic Biomarker, International Journal of Advance
Research in Engineering, Science & Technology, 2017
[5] Tanushree Sinha Roy, Neeraj Sirohi, ArtiPatle ,
”Classification of Lung Image and Nodule Detection
Using Fuzzy Inference System,” International
Conference on Computing, Communication and
Automation ,2015

IRJET- Lung Cancer Detection using Grey Level Co-Occurrence Matrix

More Related Content

What's hot (20)

Similar to IRJET- Lung Cancer Detection using Grey Level Co-Occurrence Matrix (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Lung Cancer Detection using Grey Level Co-Occurrence Matrix