CHAPTER 1

INTRODUCTION

Artificial Intelligence (AI) Deep learning is a subset of machine learning, which is

essentially a neural network with three or more layers. These neural networks attempt to
simulate the behavior of the human brain—albeit far from matching its ability—allowing it to
“learn” from large amounts of data. While a neural network with a single layer can still make
approximate predictions, additional hidden layers can help to optimize and refine for accuracy.
Deep learning drives many artificial intelligence (AI) applications and services that improve
automation, performing analytical and physical tasks without human intervention. Deep
learning technology lies behind everyday products and services (such as digital assistants,
voice-enabled TV remotes, and credit card fraud detection) as well as emerging technologies
(such as self-driving cars). Deep learning distinguishes itself from classical machine learning
by the type of data that it works with and the methods in which it learns. Machine learning
algorithms leverage structured, labeled data to make predictions—meaning that specific
features are defined from the input data for the model and organized into tables. This doesn’t
necessarily mean that it doesn’t use unstructured data; it just means that if it does, it generally
goes through some pre-processing to organize it into a structured format. Deep learning
eliminates some of data pre-processing that is typically involved with machine learning. These
algorithms can ingest and process unstructured data, like text and images, and it automates
feature extraction, removing some of the dependency on human experts.

1
 F i g 1 : V G G C o n v For example, let’s say that we had a set of photos of
different pets, and we wanted to categorize by “cat”, “dog”, “hamster”, et cetera. Deep
learning algorithms can determine which features (e.g., ears) are most important to distinguish
each animal from another. In machine learning, this hierarchy of features is established
manually by a human expert. Then, through the processes of gradient descent and
backpropagation, the deep learning algorithm adjusts and fits itself for accuracy, allowing it to
make predictions about a new photo of an animal with increased precision.
Machine learning and deep learning models are capable of different types of learning as
well, which are usually categorized as supervised learning, unsupervised learning, and
reinforcement learning. Supervised learning utilizes labeled datasets to categorize or make
predictions; this requires some kind of human intervention to label input data correctly.
In
contrast, unsupervised learning doesn’t require labeled datasets, and instead, it detects patterns
in the data, clustering them by any distinguishing characteristics. Reinforcement learning is a
process in which a model learns to become more accurate for performing an action in an
environment based on feedback in order to maximize the reward.
1.1 OBJECTIVES

• The ongoing pandemic of coronavirus disease 2019 (CANCER DISEASES (IQ-

OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED ) has led to global

2
 health and healthcare crisis, apart from the tremendous socioeconomic effects.

• One of the significant challenges in this crisis is to identify and monitor the CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED
patients quickly and efficiently to facilitate timely decisions for their treatment,
monitoring, and management.

1.2 SCOPE

Human lungs are the organs of respiration and each lung consists of pulmonary lobes which are

separated by the fissures.

Fig : 2 Classifications Layer

The fissure that separates its own pleural cover of each lung

3
 CHAPTER 2

LITERATURE SURVEY

LIME-Enabled Investigation of Convolutional Neural Network Performances in

CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS
COLLECTED Chest X-Ray Detection, (IEEE-2021) Automatic feature recognition through
the use of deep learning techniques is now widely used in biomedical applications. For medical
imaging applications, convolutional neural networks (CNNs) have proven to be extremely
effective at extracting subtle features that are not readily perceptible by humans. Many
researchers have recently published papers using CNNs to detect CANCER DISEASES (IQ-
OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED in CXR images due to two
main factors: the increasing availability of CXR image datasets and the broad access of free
machine learning (ML) software and tools. The transfer learning technique is also a major
contributor to the successful results in many such papers. Transfer learning allows researchers
to train only the final (top) layers of a very deep CNN using a relatively small dataset while
retaining the predictive power of the rest of the pre-trained network. This technique was used
by Chowdhury et al. to fine-tune several pre-trained ImageNet models to detect CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED in CXR
images. The same concept was also applied by Apostolopoulos and Mpesiana to classify
CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED
in CXR images among both healthy patients and several types of viral and bacterial
pneumonia.
Lung Tumor Classification and Detection from CT Scan Images using Deep
Convolutional Neural Networks (DCNN), IEEE-2019 This paper proposes a DCNN based
architecture for classifying the lung tumor. If the tumor is benign then chances of cancer could
be low and cancer can be initial stage and also curable. When the tumor becomes malignant
then chances of cancer to be high. Deep neural networks consisting of multiple hidden layers
such as convolutional layer, pooling layer, and fully connected layer. Convolutional layer is
characterized by local connectivity and weight sharing. This layer's nodes are grouped into
feature maps that share the same weights, weight sharing radically reduces the number of
4
network parameters and thus increases efficiency and prevents over fitting. Through
aggregating small sub-sets of values, the pooling layer is used to subsample the previous layer.
Max or average pooling substitutes input values by maximum or average values to reduce
output sensitivity to minor input changes. The classification results are generated by a fully
connected layer.

Conventional Neural Network approach for the Diagnosis of Lung Tumor, IEEE-
2020 There exist many types of cancers. Lung cancer has been one of the most common and
deadly diseases in the world. It is the leading cause of cancer deaths in both women and men
not only in developed but also in developing countries. The prognosis and the cure of lung
cancer depend highly on the early detection and treatment of small and localized tumors.
Among the five different conventional NN classifiers, it is noticed that the performance of
MLP Neural Network with quick propagation (QP) learning rule for Data-base I and Data-base
II is satisfactory as compared to the other ones. Therefore, it is inferred that the optimal neural
network classifier for the diagnosis of lung tumor should be MLP neural network with QP
learning rule when knowledge-base comprising of histogram coefficients and image statistics
parameters is used.
Deep learning to distinguish CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED from other lung infections, pleural diseases,
and lung tumors, IEEE-2020 This paper aims to use several deep learning architectures and
medical images to clearly distinguish CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED from other respiratory diseases that do have
similar symptoms. Specifically, it is the goal of this paper to differentiate CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED from
pneumonia, pleural effusion, and lung mass for sample radiographs of these diseases). To
accomplish this goal, we used six different deep learning architectures, ran CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED detection
experiments and analysed the results. We believe, this may help eliminate the false-positives
that may occur during CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET
WAS COLLECTED diagnosis. To separate CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED from other lung infections, pleural diseases and
lung tumours, we used deep learning methods on chest radiographs. Specifically, we used six
5
methods to distinguish CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET
WAS COLLECTED from pneumonia, pleural effusion and lung mass.
An Automatic Lung Cancer Detection and Classification (ALCDC) System Using
Convolutional Neural Network, IEEE-2020 A number of computer-aided tumor detection
and characterization techniques were suggested in the literature. Two key categories for the
development of these techniques are computer-aided (CAD) identification and CADx
diagnosis. The installation of a CADx device will minimize the amount of unnecessary
biopsies, reducing mental trauma in patients with benign tumors. CADx, therefore, acts as a
second approach and assists cancer diagnostics experts at the earlier stages of disease. This
diagnosis is accurate and efficient. In recent years, profound learning techniques have
demonstrated the possibility to
uncover features from training images automatically and to manipulate contact, even
hierarchy, between the characteristics of a deep neural structure. The new learning system can
also solve feature computing, selection and integration problems without using complicated
processing and pattern recognition steps.
Convolutional Neural Network approach for the Classification and Recognition of
Lung Nodules, IEEE 2020 Cancer or carcinoma is the abnormal growth of cells, which
ultimately leads to the stoppage of the essential cellular functions of the organism. These cells
clump together in the lungs to form ‘tumors. These tumors have the potential of encroaching
other parts of the body to cause a potential harm. Survival rate of lung carcinoma is
exceptionally less and concerning half the patients diagnosed with the unwellness die among
one year. Lung nodules are small masses of tissue that seem as white shadows on a CT scan.
The identification of respiratory organ nodules on an imaging test is a difficult task because of
low distinction of the lesions, double overlap with ribs and enormous respiratory organ vessels,
because of their irregular form and size
Classification and Segmentation Techniques for Detection of Lung Cancer from
CT Images, IEEE 2018 The system consists of Pre-processing stage, Segmentation stage,
Feature Extraction stage and Classification stage. In the pre-processing stage, quality of the
image is enhanced and noise from the image is removed. This is important since CT images
have noise and makes it difficult to identify the tumour region correctly. In the Segmentation
phase, the image is divided into several segments. The ultimate aim is to represent the image

6
into something valid for easier analysis. In feature extraction, the large data set is reduced into
set of features. Features are distinctive properties of the input and helps in differentiating the
categories of the input pattern. Features represent the raw image and are given to the
classification process. In classification, the images are classified according to the features
given to it. Classification forms classes and differentiates the images. The aim of classification
is not only to reach the accuracy but to identify which part of the body is infected by cancer.
Lung cancer images can be classified as cancerous or noncancerous. Several machine learning
and data mining techniques can be used for segmentation and classification purpose for
improving the accuracy of the cancer detection.
A Deep Learning Based Approach to Lung Cancer Identification, IEEE- 2018 In
the study, binary classification of the lung nodule is performed, emphasizing the early
diagnose of
lung cancer. Convolutional neural networks are preferred for processing. It is used during the
implementation of the tensorflow method, which is one of the popular libraries. Training and
testing procedures are shown using SPIE-AAPM dataset. Convolutional neural networks
(CNN) are multi-layered structures. CNN is a feed forward and very e ffective method
especially in detecting. Network structure is simple; has less training parameters. CNN is more
like biological neural networks thanks to its weight-sharing network structure. However, the
complexity of the network model and the number of weights is reduced. In the structure of the
network, there are layers named as Convolution, Activation, Pooling and fully connected layer.
CNN has a loss function, such as softmax, in the final layer, as in traditional neural networks.
Deep learning-based diagnosis of CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED using chest CT-scan images, IEEE-2020 This
paper used EfficientNet b4 to discriminate between COVID CT-scan images and normal CT-
scan images. Different learning rate policies are studied, such as reduce on plateau, cyclic
learning rate and constant learning rate. Instead of using binary cross-entropy, focal loss is
used to tackle hard examples. Previous studies claimed highest accuracy of 0.86 and F1-score
of 0.85 using this dataset. Our results with 0.90 F1 score and accuracy on the same dataset are
significantly better. The Coronavirus disease (CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED ) is an infectious disease that primarily affects
lungs. This virus has spread in almost every continent. Countries are racing to slow down the
7
spread by testing and treating patients. To diagnose the infected people, reverse transcription-
polymerase chain reaction (RT-PCR) test is used. Because of colossal demand; PCR kits are
under shortage, and to overcome this; radiographic techniques such as X-rays and CT-scan can
be used for diagnostic purpose.
Rapid CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET
WAS COLLECTED Diagnosis Using Deep Learning of the Computerized Tomography
Scans, IEEE 2020, Rapid diagnosis of CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED symptoms is of utmost importance. This paper
implemented CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS
COLLECTED detection models by using six different machine learning algorithms, as SVM,
NB, GBDT, AdaBoost, CNN and MLP based on the datasets that include image data on 980
patients suspected with CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET
WAS COLLECTED infection. The main purpose of this study was to introducing image-
based ML methods for developers and end-users of intelligent medical systems in a
comprehensive way. We compared the performance of the state-of-the-art models to show that
how important is the ML image-based models reliability to diagnose diseases. The
experimental results and discussions proved that the SVM with RBF kernel outperforms other
existing methods. Multilayer perceptron (MLP) is a well-known ANN in which neurons
are distributed in thoroughly connected layers. These layers are divided into three groups:
input layers, output layers, and hidden layers. The weighted inputs are linearly combined by
their corresponding neuron; then, the results are transferred through a nonlinear activation
function. Usually, a gradient-descent algorithm called back-propagation is used to train an
MLP. In this algorithm, a maximum error is defined to be used as a criterion to stop the
iterative weight update process.
Lung Segmentation-Based Pulmonary Disease Classification Using Deep Neural
Networks, IEEE 2021 Several approaches for the automated identification of pulmonary
disease using different imaging modalities have been proposed in the literature. These
approaches can be broadly categorized as, classification using the entire input image and
classification using segmented region of interest. The approaches that fall under the first
category extract meaningful information of texture and shape from the entire input image and
classifies it as healthy or infected. On the other hand, the approaches that fall under the second

8
category extract the meaningful texture and shape-related features from the segmented region
and classifies it as healthy or infected. The related work is divided based on the above
discussed categories. The motivation of this research study is to improve the classification
performance of pulmonary diseases by proposing a hierarchical system using CXR images.
The effects of pulmonary diseases can take many forms and they appear on the lungs as per
domain experts. These visual symptoms are overlapping for different types of pulmonary
diseases. For an accurate diagnosis, it is very important to identify these visual differences
individually associated with each type of pulmonary disease.
Deep Learning Algorithm for Classification and Prediction of Lung Cancer using
CT Scan Images, IEEE 2019 Cancer is a disease that is the second leading cause of death
globally, and most of the cancer related deaths are caused due to lung cancer. Lung Cancer is a
malignant tumor which is been characterized by uncontrollable growth of cancerous nodules in
the lung tissue. Lung cancer is a lethal disease i.e. it is very harmful and su fficient to cause
deaths. More than one million of deaths are occurring yearly due to lung cancer. The primary
aim of this work is to develop an advanced computer-aided diagnosis (CAD) system using
deep learning algorithms that will extract data from CT scan images and provide precise
diagnosis of Lung cancer. In this work, the deep learning techniques, namely - convolutional
neural networkand recurrent neural network are utilized to propose a model for lung cancer
diagnosis using CT scans and to obtain high accuracy.
Deep learning for lung Cancer detection and classification, SPRINGER 2020 This
work is to detect the cancerous lung nodules from the given input lung image and to classify
the lung cancer and its severity. To detect the location of the cancerous lung nodules, this work
uses novel Deep learning methods. Here, features are classified using Deep learning. A novel
FPSOCNN is proposed which reduces computational complexity of CNN. This work uses best
feature extraction techniques such as Histogram of oriented Gradients (HoG), wavelet
transform-based features, Local Binary Pattern (LBP), Scale Invariant Feature Transform
(SIFT) and Zernike Moment. After extracting texture, geometric, volumetric and intensity
features, Fuzzy Particle Swarm Optimization (FPSO) algorithm is applied for selecting the
best feature. An additional valuation is performed on another dataset coming from Arthi Scan
Hospital which is a real-time data set. From the experimental results, it is shown that novel
FPSOCNN performs better than other techniques.
9
 Weakly Supervised Deep Learning for Whole Slide Lung Cancer Image Analysis,
IEEE 2020 Histopathology plays an essential role in cancer diagnosis and serves as the gold
standard in many medical protocols. However, the examination of whole slide histopathology
image is always the bottleneck for in-time treatments. Especially, in some developing countries
(e.g., China), the supply of histopathologists severely falls short of demand, therefore, most of
the histopathologists have to bear the heavy burden and cancer patients cannot be treated
timely. Hence, automated analysis of histopathology is highly welcome to ease the workload
and address the issue of medical manpower shortage on underserved populations and areas.
Unlike tumor detection and segmentation tasks where a large number of pixelwise annotations
are provided for researchers to design fully supervised learning algorithms, the WSI
classification problem is much more challenging since only image-level labels are available in
most cases. In addition, tumor regions are generally sparsely distributed in the entire image,
mingling with a large proportion of noncancer regions, which results in more difficulties. In
this article, we proposed a weakly supervised learning method to address the whole slide lung
cancer image classification problem with minimum annotation effort. The exploited FCN can
efficiently generate cancer likelihood that helps to retrieve discriminative regions.
Lung Nodule Malignancy Prediction in Sequential CT Scans: Summary of ISBI
2018 Challenge, IEEE 2021 Challenges offer the opportunity to bring out the collective
talents of scientific communities that would not normally subject their algorithms for
performance comparison using a blinded-independent reference dataset. For this to occur, the
challenge needs to be accessible and open. Ideally, it fosters research towards a common goal
to address the most critical needs, which in our case is to improve clinical diagnosis in lung
cancer using sequential screening LDCT. The ISBI 2018 Lung Malignancy Prediction
Challenge provided the platform to organize this community wide challenge that was made
possible by the availability of NLST patient data. The underlying mandate of this exercise was
to encourage method development and to bring the best worldwide talents in the scientific
community to propose new approaches, particularly, in the early diagnosis of lung cancer and
to evaluate the
validity of these methods on a common set of test cases. We curated a cohort of 100 subjects
who participated in the National Lung Screening Trial and had established pathological
diagnoses. Data from 30 subjects were randomly selected for training and the remaining was

10
used for testing. Participants were evaluated based on the area under the receiver operating
characteristic curve (AUC) of nodule-wise malignancy scores generated by their algorithms on
the test set.
ML and DL based Classification model of Lung Cancer using Nodule Volume
IEEE-2020 The present solution to detect the cancer diagnosis, CT scan requires several steps
with image processing, nodule masking, classification then need to confirm cancer this process
will reduce the confidence of the finding cancer cells and finalizing cancer. And also, if the
diameter is small while masking the Scan reports may give false reports. With this model
design by using a Deep learning-based CNN class one can give the training to the machine
when getting the new results, it can analyze based on the different images and updated results
on the data set. Once build the model no need to give the train the machine while giving
images to detect cancer. and it will give the best results within a short period. And the data set
contain individual nodules have been labeled by malignant or non – malignant. This will use to
train a model accurately and classify it with less loss. CNN has multiple layer’s structure so
the classification malignant or non – malignant is accurate than normal ML models
Deep Learning-Based Long Term Mortality Prediction in the National Lung
Screening Trial, IEEE 2022 This study aims to use both 3D LDCT lung volumes and readily
available clinical data from NLST to make long-term mortality predictions, with minimum
human input. More importantly, we aim to use saliency maps to identify contributing features on CT
images that link to long-term mortality, particularly non-lung-cancer-related mortality. This can help
radiologists and clinicians diagnose pathology that might not be obvious at first glance, or help
corroborate the presence of subtle but important damage. Such an approach, highlighting neglected
features on CT imaging in lung cancer screening populations may speed up decision making for
radiologists and clinicians and help optimise patient care. The current method shows reasonable
performance in predicting long-term mortality in the NLST dataset with an AUC value of 0.73. This
illustrates the feasibility of performing mortality predictions from 3D CT scans, without handcrafted
features as demonstrated in the literature.
Internet of Medical Things_Based on Deep Learning Techniques for
Segmentation of Lung and Stroke Regions in CT Scans, IEEE 2020 This work aimed to
develop an innovative medical prediagnostic method based on Health IoT through deep
learning and fine-tuning. The method proposes that learning through fine-tuning is able to
generalize the learning to different types of CT images. The method was divided into two
11
stages; classification and segmentation of Lung and Hemorrhagic Stroke on CT images. In the
First Stage of our method, a model (classifier) was developed to classify the existence of
strokes in a CT image. With the option also to classify the existence of lungs in the CTs. In this
process, if the network identifies the object of interest in the CT, it follows the Segmentation
Process, which is the second stage of the method. This (second) stage is composed of four
models to segment the pulmonary region and the hemorrhagic stroke injury. The models
presented in this study, used deep learning combined with fine-tuning, and all the models
obtained excellent results. The proposed method proved to be superior as well as fully
automatic for the classification and segmentation of lung and hemorrhagic stroke CT images.
The comparisons in the First Stage of Results showed that the method obtained excellent
results using different models, and had the best and most effective approach for pulmonary
segmentation. The comparison in the Second Stage of Results and Discussion showed that the
Detectron2 network was able to detect and segment the stroke region without the need for fine-
tuning.
Inf-Net: Automatic CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER
DATASET WAS COLLECTED Lung Infection Segmentation From CT Images, IEEE
2020 In this paper, we have proposed a novel CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED lung CT infection segmentation network, named
Inf-Net, which utilizes an implicit reverse attention and explicit edge-attention to improve the
identification of infected regions. Moreover, we have also provided a semi-supervised
solution, Semi-Inf-Net, to alleviate the shortage of high quality labeled data. Extensive experiments on
our COVID-SemiSeg dataset and real CT volumes have demonstrated that the proposed Inf-Net and
Semi-Inf-Net outperform the cutting-edge segmentation models and advance the state-of-the-art
performances. The system has great potential to be applied in assessing the diagnosis of CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED , e.g., quantifying the
infected regions, monitoring the longitudinal disease changes, and mass screening processing. Note that
the proposed model is able to detect the objects with low intensity contrast between infections and
normal tissues.
Lung Cancer Disease Diagnosis Using Machine Learning Approach, IEEE 2020
This paper studied various relevant work done on this disease and attempted to use best
image pre-processing methods, feature extraction and deep learning mechanism to get more
accurate prediction results. Ultimately, profound neural system approach is utilized to
12
achieve more accurate precision in discovery of lung malignancy and precisely anticipate
stages. Likewise, we could demonstrate that utilization of AI can possibly fundamentally
distinguish and group low populace. A feature is region of a part with some interesting
geometric properties. The purpose of feature recognition is to mathematically extract higher
level features (manufacturing data) from lower level entities (e.g. surfaces, edges, curves). This
is a pioneering technology that extracts features and their parameters from solid models. It is
he ability to identify and group topological entities such as faces in solid model, into
functionally important features such as ribs, holes, slots. CNN identifies the relation and
connection within information and learns (or is prepared) through experience, not from
programming.
.
Convolutional Neural Network
approach for the Classification and Recognition of Lung Nodules, IEEE 2020 Cancer or
carcinoma is the abnormal growth of cells, which ultimately leads to the stoppage of the essential
cellular functions of the organism. These cells clump together in the lungs to form ‘tumors.
These tumors have the potential of encroaching other parts of the body to cause a potential harm.

13
 CHAPTER 3
EXISTING SYSTEM

In existing system Imaging plays a vital role in the diagnosis of lung cancer, with the
most common modalities including chest radiography, CT, PET, magnetic resonance imaging
(MRI), and radionuclide bone scanning, but in this work, we primarily used CT images for
analysis. X-Ray imaging will show most lung tumors, but CT is used because it is more sensitive
in finding tumor size and the presence of lymph node metastases. Efficient lung segmentation
technique helps to raise the accuracy and higher decision confidence value of any lung
abnormality identification system. Computer-aided diagnosis may serve as a second reader by
analyzing nodules, lesions or tumor and providing a malignancy estimate using computer vision
and machine learning techniques.

14
 CHAPTER 4

PROPOSED SYSTEM

To propose a system The coronavirus disease 2019 (CANCER DISEASES

(IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED ) pandemic has caused
health concerns worldwide since December 2019. From the beginning of infection, patients
will progress through different symptom stages, such as fever, dyspnea or even death.
Identifying disease progression and predicting patient outcome at an early stage helps target
treatment and resource allocation. However, there is no clear CANCER DISEASES (IQ-
OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED stage definition, and few
studies have addressed characterizing CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED progression, making the need for this study
evident. Lung abnormality is one of the common diseases in humans of all age group and this
disease may arise due to various reasons. Recently, the lung infection due to SARS-CoV-2 has
affected a larger human community globally, and due to its rapidity, the World-Health-
Organisation (WHO) declared it as pandemic disease. The CANCER DISEASES
(IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED disease has adverse
effects on the respiratory system, and the infection severity can be detected using a chosen
imaging modality. The field of medical imaging introduced CAD (Computer-Aided
Diagnostic) systems which help medical specialist to identify and categories the problem. The
lesions are produced with different body parts which cause the cancer. Such lesions are
referred to as nodule if they causes cancer, otherwise non-nodule. In the design of a CAD
system, the main task is to segment the volume of particular body part, like lungs volume
should be separated from the complete image so that we can keep our focus on the object of
interests. The next task is to separate the objects in lungs volume which are not part of lungs.
These objects are unwanted lesions. These unwanted lesions are potential nodules. The next
step is to classify the potential nodules into nodules and non-nodule. Rapid and accurate
detection of CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS
COLLECTED coronavirus is necessity of time to prevent and control of this pandemic by
timely quarantine and medical treatment in absence of any vaccine. Daily increase in cases of
CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED
15
 patient’s worldwide and limited number of available detection kits pose difficulty in
identifying the presence of disease. Therefore, at this point of time, necessity arises to look for
other alternatives. Among already existing, widely available and low-cost resources, X-ray is
frequently used imaging modality and on the other hand, deep learning techniques have
achieved state-of-the-art performances in computer-aided medical diagnosis.

Experimental results demonstrate the VGG16 and VGG 19 network implementation with the area under
the receiver operating characteristics of 0.930 and 0.90, respectively. VGG19 compete well with a
sensitivity and specificity of 80% and 100% compared to VGG16 with 98% and 88%, respectively.Hence,
the implementation of VGG16 and VGG19 with good accuracy proves that CNN used for lung carcinoma
detection leading to timely diagnosis and treatment may decrease the mortality rate.learning techniques is
proposed in this work. The proposed method is implemented in four phases, viz., data augmentation,
preprocessing, stage-I and stage-II deep network model designing. Deep network implementation in two
stages is designed to differentiate CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET
WAS COLLECTED induced pneumonia from healthy cases, bacterial and other virus induced pneumonia
on X-ray images of chest. We employed MIL, a deep learning method, using quantitative CT data to
accurately predict the disease severity of CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER
DATASET WAS COLLECTED . By utilizing an inexpensive and widely available test, our model can be
used to identify patients at high risk of disease progression in the early phase of the disease, which has
important practical implications for conducting early intervention, preventing disease progression, and
reducing mortality. We recommend that confirmed CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED patients should undergo CT screening as soon as they are
admitted to the hospital, so that physicians can use our model to determine the risk of severe disease. If the
result indicates a potential worsening of the condition of the patient, closer monitoring and early
intervention should be considered before the disease severity increases. We hope that this model would be
of some help to clinicians to better manage patients and contribute to combatting CANCER DISEASES
(IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED . In proposed methodology, we can
train the medical images related to lung diseases in terms of CT scan images.

16
 CHAPTER 5

BLOCK DIAGRAM

1.1 VGG TUMOR ANALYSIS

Fig 5.1 : Block Diagram VGG Architecture Tumor Analysis

1.2 LUNG SEGEMENTATION

Fig 5.2 : Block Diagram of Lung Segmentations

17
 CHAPTER 6

SOFTWARE IMPLEMENTATION

6.1 PYTHON:

These Python is a high-level programming language designed to be easy to read and simple to
implement. It is open source, which means it is free to use, even for commercial applications.
Python can run on Mac, Windows, and Unix systems and has also been ported to Java and .NET
virtual machines. Python is a fairly old language created by Guido Van Rossum. The design
began in the late 1980s and was first released in February 1991. Python is considered a scripting
language, like Ruby or Perl and is often used for creating Web applications and dynamic Web
content. It is also supported by a number of 2D and 3D imaging programs, enabling users to
create custom plug-ins and extensions with Python. Examples of applications that support a
Python API include GIMP, Inkscape, Blender, and Autodesk Maya.

6.2 PYTHON APPLICATIONS

1. A simple language which is easier to learn

Python has a very simple and elegant syntax. It's much easier to read and write Python
programs compared to other languages like: C++, Java, C#. Python makes programming fun
and allows you to focus on the solution rather than syntax.
If you are a newbie, it's a great choice to start your journey with Python.
2. Free and open-source
You can freely use and distribute Python, even for commercial use. Not only can you use and
distribute softwares written in it, you can even make changes to the Python's source code.
Python has a large community constantly improving it in each iteration.
3. Portability
You can move Python programs from one platform to another, and run it without any changes.
It runs seamlessly on almost all platforms including Windows, Mac OS X and Linux.
18
4. Extensible and Embeddable
Suppose an application requires high performance. You can easily combine pieces of C/C++ or
other languages with Python code.
This will give your application high performance as well as scripting capabilities which other
languages may not provide out of the box.
5. A high-level, interpreted language
Unlike C/C++, you don't have to worry about daunting tasks like memory management,
garbage collection and so on.
Likewise, when you run Python code, it automatically converts your code to the language your
computer understands. You don't need to worry about any lower-level operations.
6. Large standard libraries to solve common tasks
Python has a number of standard libraries which makes life of a programmer much easier since
you don't have to write all the code yourself. For example: Need to connect MySQL database
on a Web server? You can use MySQLdb library using import MySQLdb .
Standard libraries in Python are well tested and used by hundreds of people. So you can be sure
that it won't break your application.
7. Object-oriented
Everything in Python is an object. Object oriented programming (OOP) helps you solve a
complex problem intuitively.
With OOP, you are able to divide these complex problems into smaller sets by creating objects.

19
 CHAPTER – 7
SOFTWARE REQUIREMNT

7.1 DATA COLLECTION

Computed Tomography (CT) is taken into account in concert of the simplest strategies to diagnose the
pneumonic nodules. It uses x-rays to get structural and practical info concerning the physical body.
However, the CT image quality is influenced a lot by the radiation dose. The standard of image will
increase with the quantity of radiation dose, however within the same time, this will increase the
amount of x-rays being absorbed by the lungs..

7.1.1 IMAGE ENHANCEMENT

Here, a CT image can be input to the system. The user has to select the required lung
frame image for further processing. Then each image is resized to 256*256. Then implement a
median filter to remove noises from lung images. The median filter is a nonlinear digital
filtering technique, often used to remove noise from an image or signal. Such noise reduction is a
typical pre-processing step to improve the results of later processing (for example, edge
detection on an image). Median filtering is very widely used in digital image processing because,
under certain conditions, it preserves edges while removing noise (but see discussion below),
also having applications in signal processing. Median filtering is a nonlinear method used to
remove noise from images. It is widely used as it is very effective at removing noise while
preserving edges. It is particularly effective at removing ‘salt and pepper’ type noise. The
median filter works by moving through the image pixel by pixel, replacing each value with the
median value of neighboring pixels. The pattern of neighbors is called the "window", which
slides, pixel by pixel over the entire image pixel, over the entire image. The median is calculat ed
by first sorting all the pixel values from the window into numerical order, and then replacing the
pixel being considered with the middle (median) pixel value.

7.1.2 CONTOUR EXTRACTION

Feature learning comprises a set of algorithms to transform labeled or unlabeled data to a

new space, where it can capture the parameters and patterns of variation by disentangling the
hidden features. Features are learned through 20
supervised and unsupervised learning schemes.
Numerous unlabeled data is available in each domain, e.g., images, text data, speech, which
contain several patterns of variation that can easily be collected for feature extraction, e.g. from
preprocessed images. The task of feature extraction from unlabeled data is known as
unsupervised feature learning. In linear sparse coding, the goal is to find a decomposition in
which the hidden components are sparse, meaning that they have probability densities which are
highly peaked at zero and have heavy tails. This basically means that any given input vector can
be well represented using only a few significantly non-zero hidden coefficients.

7.1.3 TUMER CLASSIFICATION:

The classification is the final step of the system. After analyzing the structure, each
section individually evaluated for the probability of true positives. Lung diseases are classified
using Convolutional neural network algorithm. CNNs represent feed-forward neural networks
which encompass diverse combos of the convolutional layers, max pooling layers, and
completely related layers and Take advantage of spatially neighborhood correlation by way of
way of imposing a nearby connectivity pattern among neurons of adjacent layers. Convolutional
layers alternate with max pooling layers mimicking the individual of complex and clean cells in
mammalian seen cortex. A BPNN includes one or extra pairs of convolution and max pooling
layers and ultimately ends with completely related neural networks. The hierarchical structure of
BPNNs is steadily proved to be the most efficient and successful manner to analyse visible
representations. The fundamental challenge in such visual tasks is to model the intra-class
appearance and shape variation of objects. The hyper-spectral data with hundreds of spectral
channels can be illustrated as 2D curves. We can see that the curve of every class has its own
visual shape which is different from other classes, although it is relatively difficult to distinguish
some classes with human eye (e.g., gravel and self-blocking bricks). We know that BPNNs can
accomplish competitive and even better performance than human being in some visual problems,
and its capability inspires us to study the possibility of applying BPNNs for classify the disease
features. The BPNN varies in how the convolutional and max pooling layers are realized and
how the nets are trained. This network varies affording to the spectral channel size and the
number of output classes of input lung data. So our proposed work overcomes irregular
boundaries separation in Lung image classification with features extraction.
DIAGNOSIS DETAILS 21
 Cancer Diseases (IQ-OTH/NCCD) lung cancer dataset was collected is a rapidly
spreading viral disease that infects not only humans, but animals are also infected because of
this disease. The daily life of human beings, their health, and the economy of a country are
affected due to this deadly viral disease. Cancer Diseases (IQ-OTH/NCCD) lung cancer dataset
was collected is a common spreading disease, and till now, not a single country can prepare a
vaccine for CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS
COLLECTED . A clinical study of CANCER DISEASES (IQ-OTH/NCCD) LUNG CANCER
DATASET WAS COLLECTED infected patients has shown that these types of patients are
mostly infected from a lung infection after coming in contact with this disease. Chest x-ray (i.e.,
radiography) and chest CT are a more effective imaging technique for diagnosing lunge related
problems. Still, a substantial chest x-ray is a lower cost process in comparison to chest CT. In
this module, we can identify the COVID and other diseases. And also provide prescription for
affected diseases.

22
 CONCLUSION

In this system n recent times, for the diagnosis of several diseases, many Computer-Aided
Diagnosis (CAD) systems have been designed. Among many life-threatening diseases, lung cancer
is one of the leading causes of cancer-related deaths in humans worldwide. It is a malignant lung
tumor distinguished by the uncontrolled growth of cells in the tissues of the lungs. Diagnosis of
cancer is a challenging task due to the structure of cancer cells. Predicting lung cancer at its initial
stage plays a vital role in the diagnosis process and also increases the survival rate of patients.
People with lung cancer have an average survival rate ranging from 14 to 49% if they are
diagnosed in time. The current study focuses on lung cancer classification and prediction based on
Histopathological images by using effective deep learning techniques to attain better accuracy. For
the classification of lung cancer as Benign, Adenocarcinoma, or Squamous Cell Carcinoma, some
pre trained deep neural networks such as VGG-19 were used. A database of 15000
histopathological images was used in which 5000 benign tissue images and 10000 malignant lung
cancer-related images to train and test the classifier. The experimental results show that the VGG-
19 architecture can achieve an accuracy of 95%.

REFERENCE

[1] Cervantes, Eduardo Gasca, and Wai-Yip Chan. "LIME-Enabled Investigation of

Convolutional Neural Network Performances in CANCER DISEASES (IQ-OTH/NCCD) LUNG
CANCER DATASET WAS COLLECTED Chest X-Ray Detection." 2021 IEEE Canadian
Conference on Electrical and Computer Engineering (CCECE). IEEE, 2021.

[2] Mohanapriya, N., B. Kalaavathi, and T. senthil Kuamr. "Lung Tumor Classification
and Detection from CT Scan Images using Deep Convolutional Neural Networks
(DCNN)." 2019 International Conference on Computational Intelligence and Knowledge
Economy (ICCIKE). IEEE, 2019.

[3] Agrawal, Vijay L., and Sanjay V. Dudul. "Conventional neural network approach for
the diagnosis of lung tumor." 2020 International Conference on Computational Performance
Evaluation (ComPE). IEEE, 2020.

[4] Serener, Ali, and Sertan Serte. "Deep learning to distinguish Cancer Diseases (IQ-
OTH/NCCD) lung cancer dataset was collected from other lung infections, pleural diseases, and
23 (TIPTEKNO). IEEE, 2020.
lung tumors." 2020 Medical Technologies Congress
 [5] Abdul, Wadood. "An Automatic Lung Cancer Detection and Classification (ALCDC)
System Using Convolutional Neural Network." 2020 13th International Conference on
Developments in eSystems Engineering (DeSE). IEEE, 2020.

[6] Bhat, Sachin, et al. "Convolutional neural network approach for the classification and
recognition of lung nodules." 2020 4th International Conference on Electronics, Communication
and Aerospace Technology (ICECA). IEEE, 2020.

[7] Lobo, Prenitha, and Sunitha Guruprasad. "Classification and segmentation techniques
for detection of lung cancer from ct images." 2018 International Conference on Inventive
Research in Computing Applications (ICIRCA). IEEE, 2018.

[8] Cengil, Emine, and Ahmet Cinar. "A deep learning based approach to lung cancer
identification." 2018 International Conference on Artificial Intelligence and Data Processing
(IDAP). IEEE, 2018.
[9] Anwar, Talha, and Seemab Zakir. "Deep learning based diagnosis of CANCER
DISEASES (IQ-OTH/NCCD) LUNG CANCER DATASET WAS COLLECTED using chest
CT-scan images." 2020 IEEE 23rd International Multitopic Conference (INMIC). IEEE, 2020.

[10] Tabrizchi, Hamed, et al. "Rapid Cancer Diseases (IQ-OTH/NCCD) lung cancer
dataset was collected diagnosis using deep learning of the computerized tomography scans."
2020 IEEE 3rd International Conference and Workshop in Óbuda on Electrical and Power
Engineering (CANDO-EPE). IEEE, 2020.

[11] Zaidi, S. Zainab Yousuf, et al. "Lung Segmentation-Based Pulmonary Disease

Classification Using Deep Neural Networks." IEEE Access 9 (2021): 125202-125214.

[12] Mhaske, Diksha, Kannan Rajeswari, and Ruchita Tekade. "Deep learning algorithm
for classification and prediction of lung cancer using CT scan images." 2019 5th International
Conference On Computing, Communication, Control And Automation (ICCUBEA). IEEE,
2019.

[13] Asuntha, A., and Andy Srinivasan. "Deep learning for lung Cancer detection and
classification." Multimedia Tools and Applications 79.11 (2020): 7731-7762.

[14] Wang, Xi, et al. "Weakly supervised deep learning for whole slide lung cancer
image analysis." IEEE transactions on cybernetics
24 50.9 (2019): 3950-3962.
 [15] Balagurunathan, Yoganand, et al. "Lung Nodule Malignancy Prediction in
Sequential CT Scans: Summary of ISBI 2018 Challenge." IEEE Transactions on Medical
Imaging 40.12 (2021): 3748-3761.

[16] Babu, Popuri Ramesh, and Inampudi Ramesh Babu. "ML and DL based
classification model of lung cancer using nodule volume." 2020 Fourth International Conference
on I-SMAC (IoT in Social, Mobile, Analytics and Cloud)(I-SMAC). IEEE, 2020.

[17] Lu, Yaozhi, et al. "Deep Learning-Based Long Term Mortality Prediction in the
National Lung Screening Trial." IEEE Access 10 (2022): 34369-34378.

[18] Han, Tao, et al. "Internet of medical things—based on deep learning techniques for
segmentation of lung and stroke regions in CT scans." IEEE Access 8 (2020): 71117-71135.

25
 [19] Fan, Deng-Ping, et al. "Inf-net: Automatic Cancer Diseases (IQ-OTH/NCCD)
lung cancer dataset was collected lung infection segmentation from ct images." IEEE
Transactions on Medical Imaging 39.8 (2020): 2626-2637.

[20] Mukherjee, Swati, and S. U. Bohra. "Lung Cancer Disease Diagnosis Using
Machine Learning Approach." 2020 3rd International Conference on Intelligent Sustainable
Systems (ICISS). IEEE, 2020.

26