1.abstract:: Lung Cancer Desies Identification Using Ann

LUNG CANCER DESIES IDENTIFICATION USING ANN
1.Abstract:
The objective of this paper is to explore an expedient image segmentation

algorithm for medical images to curtail the physicians’ interpretation of computer
tomography (CT) scan images. Modern medical imaging modalities generate
large images that are extremely grim to analyze manually. The consequences of
segmentation algorithms rely on the exactitude and convergence time. At this
moment, there is a compelling necessity to explore and implement new
evolutionary algorithms to solve the problems associated with medical image
segmentation. Lung cancer is the frequently diagnosed cancer across the world
among men. Early detection of lung cancer navigates towards apposite treatment
to save human lives. CT is one of the modest medical imaging methods to
diagnose the lung cancer. In the present study, the performance of five
optimization algorithms, namely, k-means clustering, k-median clustering,
particle swarm optimization, inertia-weighted particle swarm optimization, and
guaranteed convergence particle swarm optimization (GCPSO), to extract the
tumor from the lung image has been implemented and analyzed. The performance
of median, adaptive median, and average filters in the preprocessing stage was
compared, and it was proved that the adaptive median filter is most suitable for
medical CT images. Furthermore, the image contrast is enhanced by using
adaptive histogram equalization. The preprocessed image with improved quality
is subject to four algorithms. The practical results are verified for 20 sample
images of the lung using MATLAB, and it was observed that the GCPSO has the
highest accuracy of 95.89%.
K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 1

2.INTRODUCTION

2.Introduction:
Lung cancer, also known as lung carcinoma, is a malignant tumor characterized

by uncontrolled growth of the cell in tissues of the lung. It is mandatory to treat
this to avoid spreading its growth by metastasis to other parts of the body. Most
cancers that start in the lung are carcinomas. The two main types are small-cell
lung carcinoma and non-small-cell lung carcinoma . Long-period tobacco
smoking is the primary factor for 85% of lung cancers . About 10–15% of cases
occur in people who have never smoked but due to air pollution, secondhand
smoking, asbestos, and radon gas. Computer tomography (CT) and radiographs
are the conventional methods to detect the presence of lung cancer. The diagnosis
is confirmed by biopsy which is usually performed by bronchoscopy or CT scan.
The cause of cancer-related death among men is mainly due to lung cancer.
Hence, it is essential to determine a new robust method to diagnose the lung
cancer at an earlier stage . For the present study, 20 lung image samples and four
algorithms have been taken for analysis. It was proved that the combination of
adaptive median filter, adaptive histogram equalization, and guaranteed
convergence particle swarm optimization- (GCPSO-) based algorithm has more
accurate results among others.
MATLAB :
Matlab is a multi-paradigm numerical computing environment and
proprietary programming language developed by MathWorks. MATLAB allows
matrix manipulations, plotting of functions and data, implementation of
algorithms. Matlab can be written in C, C++, java languages.
ANN(Artifical Neural Network):

ANN MODEL:
Yin = X1 *W1 + X2 *W2 + X3*W3+Xm*Wm (Yin is called net input)

That is Y= ∑im * Xi * Wi
Output can be calculayed as Y=F(Yin)
Neuro Fuzzy Classification :

 Neuro fuzzy is used to obtain fuzzy classification rules by learning

algorithm.
Fuzzy=(not clear)
 Fuzzy classification is a process of grouping individual, having a same
characterstics into a fuzzy set.
Histogram Equivalization:
Histogram Equivalization is a computer image processing technique used to
improve contrast in image. It can be done in 3 steps:
1. Compute the histogram of the image
2. Calculate the normalized sum of histogram
3. Transfer input image to output image
Segmentation:
Segmentation , it will divides image into pixels.
Filtering:
Is a technique of modifying a image.
Dilation :
Dilation is a basic operation in mathematical morphology.First it has been
expanded to gray scale image and to complete lattics.
Entropy:
Entropy is said as corresponding states of intencysity level in which individual
pixels can adopt.

3.OBJECTIVES
3.PROJECT OBJECTIVE

 Present available statistics on Lung Cancer.
 Present data on the early detection of lung cancer.
 Present the LCP Algorithm in the approach to lung cancer suspect.
 Present LCP research results in the treatment & prevention of Lung Cancer.
4.REQUIREMENT SPECIFICATION

4.1.SYSTEM SPECIFICATION
A System Requirements Specification (SRS) (also known as a Software Requirements
Specification) is a document or set of documentation that describes the features and
behaviour of a system or software application. It includes a variety of elements (see below)
that attempts to define the intended functionality required by the customer to satisfy their
different users.
4.1.1.HARDWARE REQUIREMENTS:
O Processor:Intel®Pentium®CPU4200@1.10Ghz
O RAM : 1GB
O HDD : 50GB and above

4.1.2.SOFTWARE REQUIREMENTS:
o Operating System: Windows 7/8/8.1/10

o Language: java
o Framework: image processing
o Editor: Matlab editor
o Library: matlab support library
4.2. Software tools and technologies used:
This projectencompasses theconcept of machinelearning and
Statistics. This project makes heavy use of Matlab Library,JDBC,etc
Few of the key concepts that are utilized in the project are discussed below.
1. Machine Learning: Machine learning is an application of artificial

intelligence (AI) that provides systems the ability to automatically learn
andimprovefromexperiencewithoutbeingexplicitly programmed. Machine
learning focuses on the development of computer programs that can access

data and use it learn for themselves.
The process of learning begins with observations or data, such as

examples, direct experience, or instruction, in order to look for patterns
in data and make better decisions in the future based on the examples that
we provide. The primary aim is to allow the computers learn
automatically without human intervention or assistance and adjust
actionsaccordingly.
After incorporate extraction, a classifier is depended upon to pick the correct class of
the characters. Subsequently, there are different classifiers which depend upon
machine learning methods, for example, support vector machines (SVM), artificial
neural network (ANN), significant conviction network (DBN), convolution neural
networks (CNN), secured Markov models (Well), and artificial neural network
(ANN) have been utilized as a bit of the character acknowledgment process.
Machine Learning can be broadly classified into two categories.

 Supervised machine learning algorithms
 unsupervised machine learning algorithms
a. Supervised machine learning algorithms can apply what has been

learned in the past to new data using labeled examples to predict
future events. Starting from the analysis of a known training dataset,
the learning algorithm produces an inferred function to make
predictions about the output values. The system is able to provide
targets for any new input after sufficient training. The learning
algorithm can also compare its output with the correct,intended
outputandfinderrorsinordertomodifythemodelaccordingly.
b. Unsupervised machine learning algorithms are used when the
information used to train is neither classified nor labeled.
Unsupervised learning studies how systems can infer a function to
describeahiddenstructurefromunlabeleddata.Thesystemdoesn’t figure
out the right output, but it explores the data and can draw inferences

from datasets to describe hidden structures fromunlabeled data

In this project, we make use of supervised machine learning algorithm

(In specific linear regression algorithm) to predict the share price as
we would have the data of share prices in the previous days and we
canuse it fortraining.
2. Costfunction:Whenbuildingalinearmodelit’ssaidthatwearetrying to
minimize the error an algorithm does making predictions, and wegot
that by choosing a function to help us measure the error also called cost
function.
Evaluation metrics for classification problems, such as accuracy, are not

useful for regression problems. Instead, we need evaluation metrics
designed for comparing continuous values, here we use Root Mean
Squared Error, of course there are others but this is one of the favorites
choice and we are going to go along with it.
We write the cost function
J(β0, β1, ……βn)=
3. Gradient Descent: Gradient descent is one of those “greatest

hits”algorithms that can offer a new perspective for solving
problems.Unfortunately,it’srarelytaughtinundergraduatecomputerscience
programs. In this post I’ll give an introduction to the gradient descent
algorithm, and walk through an example that demonstrates how gradient
descent can be used to solve machine learning problems such as linear
regression.

At a theoretical level, gradient descent is an algorithm that minimizes

functions. Given a function defined by a set of parameters, gradient
descent starts with an initial set of parameter values and iteratively
movestowardasetofparametervaluesthatminimizethefunction.This
iterative minimization is achieved using calculus, taking steps in the
negative direction of the functiongradient.
It’ssometimesdifficulttoseehowthismathematicalexplanation
translatesintoapracticalsetting,soit’shelpfultolookatanexample. The
canonical example when explaining gradient descent is linear reg
To run gradient descent on this error function, we first need to compute

its gradient. The gradient will act like a compass and always point us
downhill. To compute it, we will need to differentiate our error function.
Since our function is defined by two parameters (m and b), we will need
to compute a partial derivative for each. These derivatives work out to be
regression.
Formally, this error function looks like:
Lines that fit our data better (where better is defined by our error
function) will result in lower error values. If we minimize this function,
we will get the best line for our data. Since our error function consists of
two parameters (m and b) we can visualize it as a two-dimensional
surface. This is what it looks like for our dataset:

Each point in this two-dimensional space represents a line. The height of

the function at each point is the error value for that line. You can see that
some lines yield smaller error values than others (i.e., fit our data better).
When we run gradient descent search, we will start from some location on
this surface and move downhill to find the line withthe lowesterror.
We can then update the initial approximation to reach to the minimal and
thereby obtain the values of thecoefficients

4. Logisticregression:
In regression analysis, logistic regression is estimating the parameters of

a logistic model. More formally, a logistic model is one where the log-
odds of the probability of an event are a linear combination of
independent or predictor variables. The two possible dependent variable
values are often labelled as "0" and "1", which represent outcomes such
aspass/fail,win/lose,alive/deadorhealthy/sick.Thebinarylogistic
regression model can be generalized to more than two levels of the
dependent variable: categorical outputs with more than two values are
modelled by multinomial logistic regression, and if the multiple
categories are ordered, by ordinal logistic regression, for example the
proportional odds ordinal logisticmodel.
We generally fit a sigmoid curve using the techniques of cost function

optimization and other techniques as discussedabove.
5. SVM: Support Vector Machines (SVMs) are formulated to construct

binary classifiers. From a set of labelled training patterns, defined by: (xn,yn)
€RMX{+-1}(M:isdatadimension)and{n=1,2,…..NC}whereNc is the number of
samples per a class c. For a set of functions f : RM → {+-

1}SVMsseekthefunctionfthatallowstheminimalgeneralizationerror
The selection of the appropriate f is achieved by minimizing an upper
bound on the generalization error while maximizing the margin between
the two classes. Therefore, data are classified accordingto

Where b is a bias. The optimal

hyperplane corresponds to f(x) = 0. SV is the numbers of support
vectors which are training data whose Lagrange multipliers αi are
different to zero. The kernel k(,.. ) is any mathematical
function,whichrespectsMercer’sconditions.Forpatternrecognition,
the Radial Basis Function (RBF) kernel provides commonly the best
performances.
Here σ is user-defined. Extension of SVMs for multiclass problems

can be done through various approaches . For a C-class problem the
One- Against-All (OAA), which is the earliest multi-class
implementation, performs C binary SVMs in order to separate
iteratively each class from all the others. Since OAA requires a large
training time, the One-Against- One approach with the DDAG
decision function is commonly
used.TheDDAGperformsC(C−1)/2SVMseachofwhichseparateSSstwo
classes. Presently, the DDAG is employed to perform
multiclassSVMs.
8.Artificial Neural Network
A neural network is a machine learning model which consists of connected

layers of neurons. A neuron contains a number, the so called activation.
Connections are assigned weights, which describes the strength of the signal to
the connected neuron.
Input data is fed into the first layer, activating each input neuron to some extend.
Based on the weights and an activation function the network determines which
neurons from the next layer to activate and how strong the activation is going to
be. This so called feedforward process is continued until the output neurons are
activated. The architecture of a neural network has a huge influence on which
data it can work with and its performance.
FIGURE:ARCHITECTURE OF ARTIFICIAL NEURAL NETWORK
9.Back Propagation Network
Back propagation is a sort of oversaw learning for multi-layer nets, generally called
the summed up delta run the show. Screw up data at the yield layer is back induced
to before ones, empowering moving toward weights to these layers to be revived. It
is frequently used as planning calculation in current neural system applications. The
back propagation algorithm An early time of Neural Network was made by Warren
McCulloch and Walter Pitts in 1943 which was a computational model in setting of
Science and calculation. This model made orchestrated research which was pivoted
around the utilization of Neural Networks in Fake care. Made neural framework is
on an astoundingly crucial level a work of wide number of interconnected cells. The

procedure of cells are with the honest to goodness focus on that each cell gets an
information and drives a yield for happening cells. Each cell has a pre-described
The outline underneath is a square chart that depicts the structure and work stream
of a made simulated Neural Network. The neurons are interconnected with each
other in a serial way. The system include different covered layers depending on the
assurance of examination of commitments with the dataset.
It was made by Paul Werbos in 1974 and rediscovered energetically by Rumelhart

and Parker. Since its rediscovery, the back inciting count has been all things
considered utilized as a learning figuring in manage forward multilayer neural
frameworks.
FIGURE: BACK PROPAGATION NETWORK

Feed Forward Neural Networks
Before depicting the convolution neural system, I begin with a delineation of the
fundamental, or support forward, neural system. A more thorough treatment of
this material is given in. Consider a managed learning circumstance where we
are given a course of action of checked data {(x(i), y(i))}. Here, x(i) and y(i)
mean the features and name, independently of the ith planning delineation. At an
irregular state by then, neural systems give a strategy for addressing a brain
boggling, nonlinear limit hW(x) of our data variable x. The farthest point hW(x)
is parameterized by a weights organize W that we can tune to fit our information.
The capacity to consequently extricate area particular highlights from the basic
information gives a feasible contrasting option to the technique for hand-building
that has been the more conventional approach in machine learning and PC
vision. As I show in this proposition, the utilization of these educated highlights
in conjunction with the illustrative energy of a convolution neural network
empowers us to design a high-performing and capable model with in every way
that really matters no hand-tuning. Thusly, this work speaks to a takeoff from a
portion of the more regular techniques that have been connected to the issues of
text detection and recognition.

FIGURE:FEED FORWARD NEURAL NETWORKS
Convolution Neural Networks
Thus far in my piece, I have portrayed a portion of the current techniques

scientists have utilized for detection and recognition. As a rule, these strategies
join advanced models with keenly composed highlights for the current issue. In
this postulation, I will probably represent an elective plan that does not require
shrewd, hand-composed highlights or extremely mind boggling models fusing
huge measures of earlier learning. Unsupervised element learning procedures
introduce a feasible contrasting option to the hand-building of highlights. In our
work, we at that point coordinated these scholarly highlights into a convolution
neural network . This segment gives a short outline of both neural networks and
convolution neural network (CNN).

FIGURE : CONVOLUTION NEURAL NETWORKS

5.ANALYSIS PHASE
5.1. Problem Statement:
The Boston Scientific SCOPE Team researched bronchoscopy tools and procedures and
worked with physicians to develop areas of opportunity for Boston Scientific’s pulmonary
development team. Boston Scientific asked the team to investigate the problem space of
pulmonary lesion identification and diagnosis and develop a solution space.
Typically, the problem of system can be visualized into two modules
1) Training the system
2) Testing the system

Training the system: In this module, we train the system by using the
labeled dataset. We restrict the hand gesture symbols to 5 different
styles. We can assume each style to be a class and thus we have 5
classes.Thenameofeveryclass(thestylename)servesusasthelabelto all the
images corresponding to that class, which serve as data. Later the
features are extracted from every image in each of the classes. We then
train the dataset using a multi-class classifier such as SVM, Naive Bays
oraConvolutionNeuralNetworktoobtainaclassificationmodel.This
classification model is then used fortesting
Testing the system: In this module, we supply the test image to the
trainedmodel.Thetrainedmodelreturnstheprobabilityofthetestimage
belonging to each of the trained classes. We then design a hypothesis to
decide the class of the test image. This determines the gesture that is
present in the image. This activity is done on a real time system.
5.2. Advantages and applications of gesture recognition

system
There are several advantages ,A few of them can be listedbelow.
The main benefit of screening is a lower chance of dying from lung cancer, which accounts for
many deaths in current and former smokers. Still, it’s important to be aware that, as with any
type of screening, not everyone who gets screened will benefit. Screening with LDCT will not
find all lung cancers, and not all of the cancers that are found will be found early.
Even if a cancer is found by screening, you may still die from lung cancer. Also, LDCT often
finds things that turn out not to be cancer, but have to be checked out with more tests to know

what they are. You might need more CT scans, or invasive tests such as a lung biopsy, in which
a piece of lung tissue is removed with a needle or during surgery.
If you are at a higher risk, your doctor can explain your risk and how the ACS lung cancer
screening guideline applies to you. Your doctor can also talk with you about what happens
during screening and the best places to get the yearly screening test. Lung cancer screening is
covered by Medicare and by many private health insurance plans. Your health care team can
help you find out if your insurance will provide coverage.
Screening should only be done at facilities that have the right type of CT scanner and that have
experience in LDCT scans for lung cancer screening. The facility should also have a team of
specialists that can give patients the appropriate care and follow-up if there are abnormal results
on the scans. You might not have the right kind of facility nearby, so you may need to travel
some distance to be screened.
If you are at higher risk and should be screened, you should get a LDCT every year until you
reach the age of 74, as long as you are still in good health.

5.3. Proposed System
Matlab is a multi-paradigm numerical computing environment and

proprietary programming language developed by MathWorks. MATLAB allows
matrix manipulations, plotting of functions and data, implementation of
algorithms. Matlab can be written in C, C++, java languages.
We use every single image in the dataset for training. We propose to improve
the quality of the images in the dataset using the image processing and
computer vision techniques such as morphological operations, metric
correction etc. We later associate every image with a class label. This
ensures the dataset to contain the images along with the labels. Now we
send this to a learning algorithm and obtain a trained model.
The test image undergoes the same process of quality improvement and then
it is sent to the trained model to obtain the probability of the trained model
belonging to a particular digit. We then use it to decide the class of the
gesture via a designed hypothesis. This gives us the information about the
gesture present in theimage.
5.4. Features and objectives of the proposed algorithm:
1) It recognizes the gesture in an image that is not used in training.Thus

it can be imagined as a generalizedsystem.
2) It completes the task at a much fasterspeed
3) The algorithmic complexity is just polynomial in nature, not

exponential.
4) Thememoryusageisalsoofthepolynomialorder
5) The results obtained are efficient.

DESIGN DOCUMENT
6.1 Data Flow Diagram:

A Data flow diagram(DFD) is a graphical representation of the “flow” of
data through an information system, modelling its process aspects. A DFD is
often used as preliminary step to create an overview of the system without going
into great detail, which can later be elaborated. DFD’s can also be used for the
visualization of data processing (structured design).
A DFD shows what kind of information will be input to and output from the
system, how the data will advance through the system, and where the data will be
stored. It does not show the information about whether processes will operate in
sequence or in parallel unlike a flowchart which also shows the information.


6.2 Data flow diagram levels

6.3 FlowChart:
A Flowchart is a type of diagram that represents an algorithm, workflow or
process. The flowchart shows the steps as boxes of various kinds, and their order
by connecting the boxes with arrows. This diagrammatic representation
illustrates a solution model to a given problem. Flowcharts are used in analyzing,
designing, documenting or managing a process or program in various fields.
Flowcharts are used in designing and documenting simple processes or
programs. Like other types of diagrams, they help visualize what is going on and
thereby help understand a process, and perhaps also find less-obvious features
within the process, like flaws and bottlenecks.

6.4 Sequence Diagram:

A sequence diagram shows object interactions arranged in time sequence. It
depicts the objects and classes involved in the scenario and the sequence of
messages exchanged between the objects needed to carry out the functionality of
the scenario. Sequence diagrams are typically associated with use case
realizations in the Logical View of the system under development. Sequence
diagrams are sometimes called event diagrams or event scenarios.
A sequence diagram shows, as parallel vertical lines (lifelines), different
processes or objects that live simultaneously, and, as horizontal arrows, the
messages exchanged between them, in the order in which they occur. This allows
the specification of simple runtime scenarios in a graphical manner.


7.PROGRAM CODE
7.1 Coding:
Computer programming (often shortened to programming) is a process that leads from original
formulations of computing problem to executable programs. It involves activities such as
analysis, understanding and generically solving such problems resulting in an algorithm,
verification of requirements of the algorithms including its correctness and its resource
consumption, implementation(commonly referred to as coding) of the algorithm in a target
programming language, testing, debugging, and maintaining the source code, implementation of
the build system and the management of derived artefacts such as machine code of computer
programs. The algorithm is often only represented in human-pars able from reasoned about
using logic.
Computer programmers are those who write computer software. These jobs usually involve :
 Coding
 Debugging
 Documentation
 Integration
 Maintenance
 Requirements Analysis
 Software Architecture
 Software Testing.
7.2 Code:
7.2.1 Cancer.m
function varargout = cancer(varargin)
% CANCER M-file for cancer.fig
% CANCER, by itself, creates a new CANCER or raises the existing
% singleton*.
%
% H = CANCER returns the handle to a new CANCER or the handle to
% the existing singleton*.
%
% CANCER('CALLBACK',hObject,eventData,handles,...) calls the local
% function named CALLBACK in CANCER.M with the given input arguments.
%
% CANCER('Property','Value',...) creates a new CANCER or raises the
% existing singleton*. Starting from the left, property value pairs are
% applied to the GUI before cancer_OpeningFcn gets called. An
% unrecognized property name or invalid value makes property application
% stop. All inputs are passed to cancer_OpeningFcn via varargin.

%
% *See GUI Options on GUIDE's Tools menu. Choose "GUI allows only one
% instance to run (singleton)".
%
% See also: GUIDE, GUIDATA, GUIHANDLES
% Edit the above text to modify the response to help cancer
% Last Modified by GUIDE v2.5 27-Feb-2014 18:25:32
% Begin initialization code - DO NOT EDIT

gui_Singleton = 1;
gui_State = struct('gui_Name', mfilename, ...
'gui_Singleton', gui_Singleton, ...
'gui_OpeningFcn', @cancer_OpeningFcn, ...
'gui_OutputFcn', @cancer_OutputFcn, ...
'gui_LayoutFcn', [] , ...
'gui_Callback', []);
if nargin && ischar(varargin{1})
gui_State.gui_Callback = str2func(varargin{1});
end
if nargout
[varargout{1:nargout}] = gui_mainfcn(gui_State, varargin{:});
else
gui_mainfcn(gui_State, varargin{:});
end
% End initialization code - DO NOT EDIT
% --- Executes just before cancer is made visible.

function cancer_OpeningFcn(hObject, eventdata, handles, varargin)
% This function has no output args, see OutputFcn.
% hObject handle to figure
% eventdata reserved - to be defined in a future version of MATLAB
% handles structure with handles and user data (see GUIDATA)
% varargin command line arguments to cancer (see VARARGIN)
% Choose default command line output for cancer

handles.output = hObject;
% Update handles structure

guidata(hObject, handles);
% UIWAIT makes cancer wait for user response (see UIRESUME)

% uiwait(handles.figure1);
% --- Outputs from this function are returned to the command line.
function varargout = cancer_OutputFcn(hObject, eventdata, handles)
% varargout cell array for returning output args (see VARARGOUT);
% hObject handle to figure
% Get default command line output from handles structure

varargout{1} = handles.output;
% --- Executes on button press in pushbutton1.

function pushbutton1_Callback(hObject, eventdata, handles)

% hObject handle to pushbutton1 (see GCBO)
global I
clc
[filename, pathname] = uigetfile('*.jpg', 'Pick an Image');
if isequal(filename,0) | isequal(pathname,0)
warndlg('File is not selected');

else
I=imread(filename);
axes(handles.axes1)
imshow(I);
title 'Input Image'
end
title 'Input Lung Image'

global I
t=rgb2gray(I);
he=histeq(t);
axes(handles.axes2);
imshow(he);
title 'Histogram Equalization'
global I
%% Histogram Equalization
t=rgb2gray(I);
he=histeq(t);
%% Segmentation by thresholding
threshold = graythresh(he);
bw = im2bw(he,threshold);
imshow(bw)
title 'Segmentation by Thresholding'

global I
t=rgb2gray(I);
he=histeq(t);
%% Filter
hy = fspecial('sobel');

hx = hy';
Iy = imfilter(double(bw), hy, 'replicate');
imshow(Iy,[]),
title('Filtered Image')

global I
t=rgb2gray(I);
he=histeq(t);
%% Filter
hx = hy';
se = strel('line',11,90);
bw2 = imdilate(Iy,se);
imshow(bw2),
title('Dilated')

global I
t=rgb2gray(I);
he=histeq(t);
%% Filter
hx = hy';
BW5 = imfill(bw2,'holes');
imshow(BW5)
title 'Image Filling'


global I
% figure,imshow(I);
% title 'Input Lung Image'
t=rgb2gray(I);
he=histeq(t);
% figure,imshow(he);
% title 'Histogram Equalization'
% figure,imshow(bw)
% title 'Segmentation by Thresholding'
%% Filter
hx = hy';
%% Features Extraction
%% Contrast
disp ('Contrast Value');
cmap = contrast(BW5)
%% Entropy
disp ('Entropy Value');
E = entropy(BW5)
set(handles.edit2,'string',num2str(E));
%% GLCM Values
glcm = graycomatrix(BW5,'Offset',[2 0])

global I
% figure,imshow(I);
% title 'Input Lung Image'
t=rgb2gray(I);
he=histeq(t);
% figure,imshow(he);
% title 'Histogram Equalization'
% figure,imshow(bw)
% title 'Segmentation by Thresholding'
%% Filter
hx = hy';
% figure, imshow(Iy,[]),
% title('Filtered Image')
C=BW5;

%% mean
c1=mean(C);
%% variance
c2=var(double(C));
%% contrast
d=C;
%% energy, homogeneity, contrast

GLCM2 = graycomatrix(d,'Offset',[2 0;0 2]);
c4 = graycoprops(GLCM2,{'contrast','homogeneity','Energy'});
set(handles.edit4,'string',num2str(min(c4.Energy)));
c24= graycoprops(GLCM2,'contrast');
set(handles.edit3,'string',num2str(min(c24.Contrast)));
%% correlation
c5=corr(double(d));
c6=c5(1,:);
c7=c1;
c8=c2;
c9=[c6 c7 c8];
%% network
net = network
net.numInputs = 6
net.numLayers = 1
P = size(double(c1));
Cidx = strcmp('Cancer',c9);
T = size(double(c2));
net = newff(P,T,25);
[net,tr] = train(net,P,T);
testInputs = P(:,tr.testInd);
testTargets = T(:,tr.testInd);
out = round(sim(net,testInputs));
diff = [testTargets - 2*out];
detections = length(find(diff==-1))
false_positives = length(find(diff==1))
true_positives = length(find(diff==0))
false_alarms = length(find(diff==-2))
Nt = size(testInputs,2);
fprintf('Total testing samples: %d\n', Nt);
cm = [detections false_positives; false_alarms true_positives]
cm_p = (cm ./ Nt) .* 100 ;
%%
sim_out = round(sim(net,testInputs));
if ((max(c24.Contrast))>2)
set(handles.edit1,'string','Lung Cancer Affected Image');
else
set(handles.edit1,'string','Normal Image');
end
function edit1_Callback(hObject, eventdata, handles)

% hObject handle to edit1 (see GCBO)
% Hints: get(hObject,'String') returns contents of edit1 as text

% str2double(get(hObject,'String')) returns contents of edit1 as a
double

% --- Executes during object creation, after setting all properties.

function edit1_CreateFcn(hObject, eventdata, handles)
% handles empty - handles not created until after all CreateFcns called
% Hint: edit controls usually have a white background on Windows.

% See ISPC and COMPUTER.
if ispc && isequal(get(hObject,'BackgroundColor'),
get(0,'defaultUicontrolBackgroundColor'))
set(hObject,'BackgroundColor','white');
end


double


end


double



end


double


end


double


end



double


end


double


end

8.TESTING
8.1 TESTING
The goal of utilizing numerous testing methodologies in the development process is
to make sure the software can successfully operate in multiple environments and across
different platforms. These can typically be broken down between functional and non-
functional testing. Functional testing involves testing the application against the
business requirements. It incorporates all test types designed to guarantee each part of a
piece of software behaves as expected by using uses cases provided by the design team
or business analyst. These testing methods are usually conducted in order and include:
 Unit testing
 Integration testing
 System testing
 Acceptance testing
Non-functional testing methods incorporate all test types focused on the operational
aspects of a piece of software. These include:

 Performance testing
 Security testing
 Usability testing
 Compatibility testing
8.1.1 Unit Testing
Unit testing is the first level of testing and is often performed by the developers
themselves. It is the process of ensuring individual components of a piece of software
at the code level are functional and work as they were designed to. Developers in a
test-driven environment will typically write and run the tests prior to the software or
feature being passed over to the test team. Unit testing can be conducted manually, but
automating the process will speed up delivery cycles and expand test coverage. Unit
testing will also make debugging easier because finding issues earlier means they take
less time to fix than if they were discovered later in the testing process. TestLeft is a
tool that allows advanced testers and developers to shift left with the fastest test
automation tool embedded in any IDE.
8.1.2 Integration Testing
After each unit is thoroughly tested, it is integrated with other units to create modules
or components that are designed to perform specific tasks or activities. These are then
tested as group through integration testing to ensure whole segments of an application
behave as expected (i.e, the interactions between units are seamless). These tests are
often framed by user scenarios, such as logging into an application or opening files.
Integrated tests can be conducted by either developers or independent testers and are
usually comprised of a combination of automated functional and manual tests.
8.1.3 System Testing

System testing is a black box testing method used to evaluate the completed and
integrated system, as a whole, to ensure it meets specified requirements. The
functionality of the software is tested from end-to-end and is typically conducted by a

separate testing team than the development team before the product is pushed into
production.
8.1.4 Acceptance Testing

Acceptance testing is the last phase of functional testing and is used to assess whether
or not the final piece of software is ready for delivery. It involves ensuring that the
product is in compliance with all of the original business criteria and that it meets the
end user’s needs. This requires the product be tested both internally and externally,
meaning you’ll need to get it into the hands of your end users for beta testing along
with those of your QA team. Beta testing is key to getting real feedback from potential
customers and can address any final usability concerns.
8.1.5 Performance Testing

Performance testing is a non-functional testing technique used to determine how an
application will behave under various conditions. The goal is to test its responsiveness
and stability in real user situations. Performance testing can be broken down into four
types:
 Load testing is the process of putting increasing amounts of simulated demand on

your software, application, or website to verify whether or not it can handle what it’s
designed to handle.
 Stress testing takes this a step further and is used to gauge how your software will
respond at or beyond its peak load. The goal of stress testing is to overload the
application on purpose until it breaks by applying both realistic and unrealistic load
scenarios. With stress testing, you’ll be able to find the failure point of your piece of
software.
 Endurance testing, also known as soak testing, is used to analyze the behavior of
an application under a specific amount of simulated loaSd over longer amounts of time.
The goal is to understand how your system will behave under sustained use, making it
a longer process than load or stress testing (which are designed to end after a few
hours). A critical piece of endurance testing is that it helps uncover memory leaks.
 Spike testing is a type of load test used to determine how your software will
respond to substantially larger bursts of concurrent user or system activity over varying

amounts of time. Ideally, this will help you understand what will happen when the load
is suddenly and drastically increased.
8.1.6 Security Testing

With the rise of cloud-based testing platforms and cyber attacks, there is a growing
concern and need for the security of data being used and stored in software. Security
testing is a non-functional software testing technique used to determine if the
information and data in a system is protected. The goal is to purposefully find
loopholes and security risks in the system that could result in unauthorized access to or
the loss of information by probing the application for weaknesses. There are multiple
types of this testing method, each of which aimed at verifying six basic principles of
security:
1. Integrity
2. Confidentiality
3. Authentication
4. Authorization
5. Availability
6. Non-repudiation
8.1.7 Usability Testing

Usability testing is a testing method that measures an application’s ease-of-use from
the end-user perspective and is often performed during the system or acceptance testing
stages. The goal is to determine whether or not the visible design and aesthetics of an
application meet the intended workflow for various processes, such as logging into an
application. Usability testing is a great way for teams to review separate functions, or
the system as a whole, is intuitive to use.
8.1.8 CompatibilityTesting
Compatibility testing is used to gauge how an application or piece of software will
work in different environments. It is used to check that your product is compatible with
multiple operating systems, platforms, browsers, or resolution configurations. The goal
is to ensure that your software’s functionality is consistently supported across any
environment you expect your end users to be using.

8.1.9 Testing With TestComplete
TestComplete is our robust automated GUI testing tool that excels in compatibility
and integration testing. It helps QA teams create and run tests across desktop, mobile,
and web applications – enabling testing professionals to speed up delivery cycles and
improve software quality. Testcomplete comes with built-in support for various test
environments, integrations to performance testing tools, as well as support for
developer friendly SCMs, allowing you to seamlessness integrate it into your
development process. Using TestComplete will enable you to build a robust testing
framework that utilizes the broad spectrum of available software testing
methodologies.

9.INPUT AND OUTPUT SNAPSHOTS
INPUT :
10.CONCLUSION AND FUTURE SCOPE


10.1Future Scope:
This research mainly deals with Computer Aided Diagnosis System (CAD) for early detection
of lung cancer nodules from the Chest Computer Tomography (CT) image using FPCM,
MFPCM and SVM and MFPCM and ELM approaches. These techniques are very much useful
for the detection of 135 lung cancer. The proposed CAD system with the ELM approach
provides the best result among the three proposed techniques. In order to improve the
performance of the proposed approaches, some future enhancements would be necessary in the
present research work. The main aim of the future enhancements would be to increase the
sensitivity and specificity of the system. The future option for this research is to use Enhanced
Extreme Learning Machine in the CAD system. The use of Enhanced ELM (Guang-Bin Huang
et al 2008) would provide even better results. Moreover the training time and the accuracy of
the system would also be better by using Enhanced ELM. The second future enhancement of
this research would be to use the new clustering algorithm which can provide better results than
the proposed clustering algorithms. The other future enhancements would be to incorporate
latest technologies like nanotechnology (Medley et al 2008), genetic algorithm (Serhat 2007)
etc., into the CAD system for the better performance of the system.
10.2 DISADVANTAGES:
 The patient must be completely immobile
 Dental amalgam, metallic structure,and prostheses greatly degrade the images
10.3 Conclusion:
Lung cancer is the leading cause of death over the world and the only chance of
cure for patients affected from this kind of cancer is surgical resection.
This is mainly due to the fact that several factors are involved in lung cancer
develoment and progression and to date the diagnostic methods available for an
early and efficient detection are not sufficient.
Although lung cancer research data have accumulated dramatically during the
past several years, there is no database specifically focusing on lung cancer
molecular biology available yet.

In this short review we have summarized some of the factors contributing to lung
cancer, but it is just an overview of the most important proteins involved in lung
cancer disease, which are often mutated or present an unusual pattern of
expression compared to the healthy tissue.
All together these data are important for understanding the nature behind this type
of cancer and also to understand that basic research on proteins, miRNAs and all
the other epigenetic modifications could be used to develop more powerful
diagnostic tools, as well as prognostic or predictive markers or even for the
development of new targeting therapies for lung cancer.

References:
1. Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer statistics, 2007. CA Cancer J.
Clin. 2007;57(1):43–66. [PubMed] [Google Scholar]
2. Travis WD, Travis LB, Devesa SS. Lung cancer. Cancer. 1995;75(1 Suppl):191–202. [PubMed] [Google
Scholar]
3. Schreiber G, McCrory DC. Performance characteristics of different modalities for diagnosis of suspected
lung cancer: summary of published evidence. Chest. 2003;123(1 Suppl):115S–128S. [PubMed] [Google
Scholar]
4. Henschke C I, Yankelevitz D F. CT screening for lung cancer: update 2007. Oncologist. 2008;13(1):65–
78. [PubMed] [Google Scholar]
5. Spira A, Beane J, Shah V, Liu G, Schembri F, Yang X, Palma J, Brody JS. Effects of cigarette smoke on
the human airway epithelial cell transcriptome. Proc. Natl. Acad. Sci. USA. 2004;101(27):10143–
10148. [PMC free article] [PubMed] [Google Scholar]
6. Sandler A, Gray R, Perry MC, Brahmer J, Schiller JH, Dowlati A, Lilenbaum R, Johnson DH. Paclitaxel-
carboplatin alone or with bevacizumab for non-small-cell lung cancer. N. Engl. J. Med. 2006;355(24):2542–
7. Sone S, Li F, Yang ZG, Honda T, Maruyama Y, Takashima S, Hasegawa M, Kawakami S, Kubo K,
Haniuda M, Yamanda T. Results of three-year mass screening programme for lung cancer using mobile low-
dose spiral computed tomography scanner. Br. J. Cancer. 2001;84(1):25–32. [PMC free
article] [PubMed] [Google Scholar]
8. Fossella FV, DeVore R, Kerr RN, Crawford J, Natale RR, Dunphy F, Kalman L, Miller V, Lee JS, Moore
M, Gandara D, Karp D, Vokes E, Kris M, Kim Y, Gamza F, Hammershaimb L. Randomized phase III trial
of docetaxel versus vinorelbine or ifosfamide in patients with advanced non-small-cell lung cancer
previously treated with platinum-containing chemotherapy regimens. The TAX 320 Non-Small Cell Lung
Cancer Study Group. J. Clin. Oncol. 2000;18(12):2354–2362. [PubMed] [Google Scholar]
9. Wang L, Xiong Y, Sun Y, Fang Z, Li L, Ji H, Shi T. HLungDB: an integrated database of human lung
cancer research. Nucleic Acids Res. 2010;38(Database issue):D665–9. [PMC free
article] [PubMed] [Google Scholar]
10. Fong KM, Sekido Y, Gazdar AF, Minna JD. Lung cancer. 9: Molecular biology of lung cancer: clinical
implications. Thorax. 2003;58(10):892–900. [PMC free article] [PubMed] [Google Scholar]
11. Herbst RS, Heymach JV, Lippman SM. Lung cancer. N. Engl. J. Med. 2008;359(13):1367–
12. Aviel-Ronen S, Blackhall FH, Shepherd FA, Tsao MS. Kras mutations in non-small-cell lung
carcinoma: a review. Clin. Lung Cancer. 2006;8(1):30–38. [PubMed] [Google Scholar]
13. Reissmann PT, Koga H, Figlin RA, Holmes EC, Slamon DJ. Amplification and overexpression of the
cyclin D1 and epidermal growth factor receptor genes in non-small-cell lung cancer. Lung Cancer Study
Group. J. Cancer Res. Clin. Oncol. 1999;125(2):61–70. [PubMed] [Google Scholar]
14. Salgia R, Skarin AT. Molecular abnormalities in lung cancer. J. Clin. Oncol. 1998;16(3):1207–
15. Shigematsu H, Takahashi T, Nomura M, Majmudar K, Suzuki M, Lee H, Wistuba II, Fong KM,
Toyooka S, Shimizu N, Fujisawa T, Minna JD, Gazdar AF. Somatic mutations of the HER2 kinase domain
in lung adenocarcinomas. Cancer Res. 2005;65(5):1642–1646. [PubMed] [Google Scholar]
16. Buttitta F, Barassi F, Fresu G, Felicioni L, Chella A, Paolizzi D, Lattanzio G, Salvatore S, Camplese PP,
Rosini S, Iarussi T, Mucilli F, Sacco R, Mezzetti A, Marchetti A. Mutational analysis of the HER2 gene in
lung tumors from Caucasian patients: mutations are mainly present in adenocarcinomas with
bronchioloalveolar features. Int. J. Cancer. 2006;119:2586–2591. [PubMed] [Google Scholar]

17. Naoki K, Chen TH, Richards WG, Sugarbaker DJ, Meyerson M. Missense mutations of the BRAF gene
in human lung adenocarcinoma. Cancer Res. 2002;62(23):7001–7003. [PubMed] [Google Scholar]
18. Brose MS, Volpe P, Feldman M, Kumar M, Rishi I, Gerrero R, Einhorn E, Herlyn M, Minna J,
Nicholson A, Roth JA, Albelda SM, Davies H, Cox C, Brignell G, Stephens P, Futreal PA, Wooster R,
Stratton MR, Weber BL. BRAF and RAS mutations in human lung cancer and melanoma. Cancer
Res. 2002;62(23):6997–7000. [PubMed] [Google Scholar]
19. Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto RA, Brannigan BW, Harris PL, Haserlat
SM, Supko JG, Haluska FG, Louis DN, Christiani DC, Settleman J, Haber DA. Activating mutations in the
epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N.
Engl. J. Med. 2004;350:2129–2139. [PubMed] [Google Scholar]
20. Paez JG, Jänne PA, Lee JC, Tracy S, Greulich H, Gabriel S, Herman P, Kaye FJ, Lindeman N, Boggon
TJ, Naoki K, Sasaki H, Fujii Y, Eck MJ, Sellers WR, Johnson BE, Meyerson M. EGFRmutations in lung
cancer: correlation with clinical response to gefitinib therapy. Science. 2004;304:1497–

1.abstract:: Lung Cancer Desies Identification Using Ann

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1.abstract:: Lung Cancer Desies Identification Using Ann

Uploaded by

Copyright:

Available Formats

LUNG CANCER DESIES IDENTIFICATION USING ANN

The objective of this paper is to explore an expedient image segmentation

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 1

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 2

Lung cancer, also known as lung carcinoma, is a malignant tumor characterized

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 3

Yin = X1 *W1 + X2 *W2 + X3*W3+Xm*Wm (Yin is called net input)

Neuro Fuzzy Classification :

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 4

 Neuro fuzzy is used to obtain fuzzy classification rules by learning

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 5

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 6

 Present available statistics on Lung Cancer.

 Present data on the early detection of lung cancer.

 Present the LCP Algorithm in the approach to lung cancer suspect.

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 7

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 8

o Operating System: Windows 7/8/8.1/10

4.2. Software tools and technologies used:

This projectencompasses theconcept of machinelearning and

Statistics. This project makes heavy use of Matlab Library,JDBC,etc

1. Machine Learning: Machine learning is an application of artificial

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 9

data and use it learn for themselves.

The process of learning begins with observations or data, such as

Machine Learning can be broadly classified into two categories.

a. Supervised machine learning algorithms can apply what has been

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 10

from datasets to describe hidden structures fromunlabeled data

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 11

In this project, we make use of supervised machine learning algorithm

Evaluation metrics for classification problems, such as accuracy, are not

J(β0, β1, ……βn)=

3. Gradient Descent: Gradient descent is one of those “greatest

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 12

At a theoretical level, gradient descent is an algorithm that minimizes

To run gradient descent on this error function, we first need to compute

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 13

Each point in this two-dimensional space represents a line. The height of

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 14

In regression analysis, logistic regression is estimating the parameters of

We generally fit a sigmoid curve using the techniques of cost function

5. SVM: Support Vector Machines (SVMs) are formulated to construct

€RMX{+-1}(M:isdatadimension)and{n=1,2,…..NC}whereNc is the number of

samples per a class c. For a set of functions f : RM → {+-

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 15

Where b is a bias. The optimal

Here σ is user-defined. Extension of SVMs for multiclass problems

8.Artificial Neural Network

A neural network is a machine learning model which consists of connected

FIGURE:ARCHITECTURE OF ARTIFICIAL NEURAL NETWORK

9.Back Propagation Network

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 17

It was made by Paul Werbos in 1974 and rediscovered energetically by Rumelhart

FIGURE: BACK PROPAGATION NETWORK

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 18

Feed Forward Neural Networks

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 19

FIGURE:FEED FORWARD NEURAL NETWORKS

Convolution Neural Networks

Thus far in my piece, I have portrayed a portion of the current techniques

K.L.E’s Smt C.I.MUNAVALLI POLYTECHNIC HUBBALLI Page 20

FIGURE : CONVOLUTION NEURAL NETWORKS

Yin = X1 W1 + X2 W2 + X3W3+XmWm (Yin is called net input)