Technical Seminar Presentation

On
LUNG CANCER PREDICTION USING
CONVOLUTIONAL NEURAL NETWORK.
By

Teja D 4GL19EC042

Under the guidance of

Prof. MANJUNATHA A Dr. A L CHOODARATHNAKARA

Asst. Professor in Department of E&C HOD in Department of E&C
Government Engineering College Government Engineering College
Kushalnagar - 571234 Kushalnagar - 571234
 Table Of Content

• Introduction
• Relevance and significance
• Objectives
• Literature review
• Methodology
• Convolutional Neural Network
• Advantages
• Results
• Conclusion
• References
 INTRODUCTION
• The abnormal growth of cells in human Lung is called as Lung Cancer.

• The extra cells can divide without stopping and may form growths called tumor.
Uncontrolled magnification of tissue creates lung cancer.
• Lung cancer is one of the most serious diseases in the world today, and it is the
leading cause of mortality in the previous several decades.
• The addiction to cigarettes is one of the leading causes of lung cancer.

• Furthermore, carcinogenic surroundings such as radioactive gas and air pollution

contribute to the spread of this disease. In addition, genetic factors also have a
major contribution to lung cancer.
• If it is detected and identified in primary stage then survival rate of many
numbers of patients can be improved.
• Later after disease identification, by providing proper diagnosis can reduce the
death rate of patients.
• Training deep neural networks to detect lung cancer.

• In the process of lung cancer prediction, the images applied at the input layer of
Deep Convolutional neural networks are classified into cancerous or non-
cancerous.
Relevance and significance
• The Machine Learning model is built to predict the lung cancer using the LIDC
(lung image database consortium),IDRI(image database resource initiative)
dataset .
• The CNN algorithm is used to build the prediction model .

• In the process of lung cancer classification, the images applied at the input layer
of Deep Convolutional neural networks are classified into cancerous or non-
cancerous .
Objectives
• Benefit of early-detection of lung cancer.

• To find the Machine learning model which gives the low Mean absolute error and
high accuracy.
• The objective is to reduce the mortality(death) rate of lung cancer patients.
Literature Review

• The Convolutional neural network was used for classification by using images of
dataset, 97% of accuracy results found with 90% training and 10% of testing
images.
• Identification of lung nodules by applying computed tomography images.

• The method to classify the lung nodule by computed tomography images where
the lung segmentation take place, thereby the image features are extracted.
• The extracted features were used to apply as input to the various classifiers, then
the classifier decides and detection of cancer.
• Deep Convolutional neural network is trained from CT images of LIDC dataset to
predict the lung cancer.
• The Convolutional neural network classifier for identifying lung nodules which
gives an accuracy of about 97%.
• It is noted that the degree of treating the diseases will be higher as the dataset
quantity increases.
• The model which is used to identify cancerous part of the lung by applying
the methods of deep learning of neural network.
• The system uses the methods of machine learning such as K-Nearest
Neighbor, Random Forest; the system gives an accuracy of classification of
82%.
• SVM classifier gives an accuracy of 86%. Compare to these results of
classification, CNN provides the more accuracy then SVM classifier.
• It is evident from the review that one of the most powerful tools to classify the
cancerous images is Convolutional neural networks and its deep learning features.

ALGORITHM ACCURACY
Support vector machine 86%
(SVM)

Random forest, KNN 82%

Convolutional neural 97%
network (CNN)
Methodology:
• Collection of LIDC-IDRI dataset.

• Preprocessing the collected dataset.

• Splitting dataset into training data (90%) and testing (10%).

• Building a prediction model.

• Obtaining the results.

Convolutional Neural Network:
1. Convolutional layer: This layer accepts the input images of specified size
suitable for the network training, which is then, translated into feature maps
by using filters or Convolutional kernels. The filters used in this layer are
moved through the dimensions.

2. Pooling layer: The important function of this layer is to downsize the matrix
and to minimize the parameters hence this layer does down-sampling from
the feature maps of Convolutional layers. This layer calculates maximum
value or weighted average by moving filters across the output of
Convolutional layer.
• Fully-connected layer: The objective of this layer is to classify the resulting
images of the previous two layers into a label. Since this layer utilizes the
SoftMax layer so as to find the probabilities of values in between 0 and 1. Besides
it uses Batch normalization to enhance the training rate and to minimize over
fitting. The Identification of lung cancer using Deep CNN, comprise of two
categories. The first category does pre-process functionalities suitable to train and
process the images in DCNN, thus feature extraction can be performed and
second category performs the classification of input CT images where it identifies
the cancer in lungs.
Fig. Input Dataset
Fig.Input CT scan image Fig.Preprocessed CT scan

Prediction: cancer
Fig. Input CT scan image Fig.Preprocessed CT scan

Prediction: non-cancer
Advantages:

• Helps for diagnosing the patient in the early stages.

• Before symptoms start helps lower the risk of dying from lung
cancer.
• Prediction is fast and achieve good accuracy.
Results:

• Training and testing images are categorized for evaluating the network for
efficient classification of images into cancerous and non-cancerous images and
helps for diagnosing the patient in the early stages.
• The CNN model achieves a good accuracy of 97%.
Fig.Input Dataset for proposed method
Prediction : cancerous
Prediction : non-cancerous
 CONCLUSION
• It can be seen that convolutional neural network is providing more
accuracy compared to other algorithms.
• This method is faster, more accurate and more efficient than the
other technique.
• This is advantageous for predicting whether a patient will develop
cancer or not.
Bibliography:

[1] X. Huang, J. Shan, V. Vaidya, Lung Nodule Detection in CT Using 3D Convolutional

Neural Networks, IEEE, 2017.

[2] X. Yue, et al., Cloud-assisted Industrial Cyber-Physical Systems: an Insight, 2015.

[3] M.D. Zeiler, R. Fergus, Visualizing and understanding convolutional networks, in:
Proceedings of the Europea Cconference on Computer Vision, ECCV), 2014, 818–8ss.

[4] R. Sindhu, S. Jose George, S. Skaria, V.V. Varun, Using YOLO Based Deep Learning
Network for Real Time Detection and Localiza Tion of Lung Nodules from Low Dose CT
scans, 2018.

[5] Y. LeCun, L. Bottou, Y. Bengio, P. Haffner, Gradient based learning applied to

document recognition, Proc. IEEE 86 (11) (1998) 2278–2324.
[6] T. Barua, R. Doshi, K.K. Hiran, Mobile Applications Develop Ment with python in
Kivy Framework, 9783110689389 De Gruyter Stem, 2021, ISBN 3110689383.

[7] Kaggle, Kaggle data science bowl 2017. https://www.kaggle.com/c/data-science

bowl-2017, 2017. (Accessed 20 December 2020).

[8] K. Simonyan, A. Zisserman, Very deep convolutional networks for large-scale image
recognition, in: Proceedings of the International Conference on Learning Representations
(ICLR), 2015.

[9] R. Van Kranenburg, A Critique of Ambient Technology and the All-Seeing Network of
RFID, The Netherlands Institute of Network Culture, Amsterdam, 2008.

[10] E. Romera, J.M. Alvarez, L.M. Bergasa, R. Arroyo, Efficient Con vNet for Real-
Time Semantic Segmentation, IEEE, 2017.
THANK YOU