Lung cancer prediction using deep learning
2023-2024
DEPARTMENT OF INFORMATION TECHNLOGY
KAVIKULGURU INSTITUTE OF TECHNOLOGY AND SCIENCE
RAMTEK, NAGPUR, MAHARASHTRA, INDIA-441 106
Presented by
Ms. Kantima Tiwari (IT20030)
Ms. Shrutuja Bawankule (IT20027)
Under the guidance of
Ms. Geeta Meshram (IT20010) Mrs. Saroj Shambharkar
Assistant Professor & Head of the Department
Ms. Isha Chilkuri (IT20063)
Ms. Sona Gajbiye (ITD20088)
CONTENTS
Introduction
Abstract
Literature Survey
Architecture
Drawbacks of Existing System
Introduction
Prediction of lung cancer using deep learning model ,In this , we will be predicting
the lung cancer in human Body based on x- ray images .Based on the x ray images
our system will extract features
From it and will tell us about the stage of lung cancer Lung cancer occurs due the
growth of cancerous cell In the lung. Which destroy the that particular organ
In the body there is one more type pneumonia which
is considered as a lung cancer or lung disease .
There Are various reasons of lung cancer such as Smoking and consumption of
alcohol.
In this project we are going to develop a model for the prediction of stages of lung
cancer or the growth of lung cancer by using that model .
We are also focusing on the accuracy rate of out model for best prediction of lung
cancer stages.
ABSTRACT
Lung cancer is increasing rapidly due to various reasons such as smoking but the main
cause that can affect lung of non smoking person is air pollution.
It is very important to know the level or stage of lung cancer , if we are able to detect
the lung cancer at first or initial stage there are very high chances that the patient can be
saved .
For this purpose we are going to use deep learning concepts to develop a system which
is able to do the accurate prediction of lung cancer based on some features
It is very important to detect the disease at initial state . We are going to use CT scan
images or X-ray images as input dataset for lung cancer detection neural network is also
used in this project
We are going to use convolutional neural network in this prediction , if we are able to
detect the lung cancer at initial level the survival rate of patient can be increased .
KEYWORD
Lung cancer , Deep Learning , CT scan images , Neural network
LITERATURE SURVEY
Year of Research Input / Approach/ Drawback Outcome of
publications paper title & dataset used algorithm their work
publication by them technique
detail
April 2020 Prediction of lung CT(Computed AI(CNN,ANN, the CT filter picture is The proposed model
cancer using deep tomography) scan Deep learning, preprepared and is trained using the
learning algorithm images Feed forward back pursued by division of trained dataset and
Parthasarathy G, propagation) the ROI of the lung, the unknown values
2Abirami S, 3Monica the are predicted
Santhana A,4Nishali Discrete waveform
C, 5Pavithrasrisai V Transform is
1,2,3,4,5SRM TRP connected for picture
Engineering College, pressure and
Tiruchirappalli, India highlights are
extricated utilizing a
GLCM
January 2022 Deep learning based CT(Computed CNN, Deep Learning model was developed DL-based model
algorithm for lung tomography) scan using a dataset developed using the
cancer detection images collected from a single segmentation method
Akitoshi Shimazaki, hospital. included showed high
Daiju Ueda1,2, only chest radiographs performance in the
Antoine Choppin3 , containing malignant detection of lung
AkiraYamamoto1 , nodules/masses cancer on chest
Takashi Honjo1 , used only radiographs
Yuki Shimahara3 & pathologically proven
Yukio Miki lung cancers a
Year of Research Input / dataset Approach/ Drawback Outcome of
publications paper title & used by them algorithm their work
publication technique
detail
2023 B. S. Deepapriya, CT scan Deep learning Lung Chest radiographs on Computer-Aided
Parasuraman Kumar, cancer detection which radiologist diagnosis is
G. Nandakumar, S. could not identify the considered to bring a
Gnanavel, R. Neural networks lesion even with the boost to the field of
Padmanabhan, Transfer learning reference to ct medicine by tying it to
Anbarasa Kumar automated systems
Anbarasan, K. Meena
“Performance
evaluation of deep
learning techniques
for lung cancer
prediction”
November 2022 Early prediction of CT Scan or chest X-ray Deep learning , CNN Segmentation of lung They have did correct
lung cancer using images structure is very prediction for 50%
deep saliency capsule challenging problem normal,50%
and pre-trained deep because there is no malingnant,50% &
learning framework homogeneity in lung 50%benign
region
DIAGRAM TO ILLUSTRATE PROPOSED WORK
ARCHITECTURE / PROPOSED WORKFLOW
In this , we are creating a project for lung cancer prediction , for that we are using image
dataset as input .
• Image Dataset(Input): Image datasets help algorithms learn to identify and recognize
information in images and perform related cognitive activities. For example, AI
algorithms can be trained to tag photographs, read car plates, and identify tumour’s in
medical images.
• Deep learning : The project is based on Deep Learning. Deep learning uses artificial
neural networks to perform sophisticated computations on large amounts of data. It is a
type of machine learning that works based on the structure and function of the human
brain.
• Some steps are as follows which are showing the flow of our proposed work system
Step 1 : import libraries & dataset
Step 2 : Data visualization – visualize some image to build classifier
Step 3 : image data processing – convert the images into NumPy arrays
Step 4 : Model development – we will be developing a simple CNN model which
will be containing some layers such as convolutional layer, flatten layer to flatten
the output of the convolutional layer, fully connected layer ,batch normalization to
enable the stable and fast learning output layer .Dropout layer before the final
layer to avoid any possibility of overfitting.
DRAWBACK OF EXSISTING SYSTEM
• In some cases, the application still does not have accurate results. Further
optimization is needed.
• Priority information is needed for segmentation, Database extension is
required for greater accuracy.
• Possible causes that can cause misclassification can be :Symptoms of the
disease vary from cigarettes, optimizing the characteristics needed, more
training pattern are needed to cover and predict more cases- the actual
disease.