VISVESVARAYA TECHNOLOGICAL UNIVERSITY
JNANA SANGAMA, BELAGAVI- 590018
Dissertation
on
“Bone Cancer Detection And Classification Using CNN”
Submitted in Partial Fulfillment for the Award of Degree of
Master of Technology
In
VLSI DESIGN AND EMBEDDED SYSTEM
Submitted by
Manoj P
1RN19LVS02
Carried Out under the Guidance
of
Dr. Manjula V K
Associate Professor
DEPARTMENT OF ELECTRONICS AND COMMUNICATION ENGINEERING
(Accredited by NBA for the Academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
RNS INSTITUTE OF TECHNOLOGY
(AICTE Approved, VTUAffiliated and NAAC ‘A’ Accredited)
(UG programs – CSE, ECE, ISE, EIE and EEE have been accredited by NBA
for the Academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
Channasandra, Dr. Vishnuvardhan Road, Bengaluru – 560 098
2020 – 2021
RNS INSTITUTE OF TECHNOLOGY
(AICTE Approved, VTUAffiliated and NAAC ‘A’ Accredited)
(UG programs – CSE, ECE, ISE, EIE and EEE have been accredited by NBA
for the Academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
Channasandra, Dr. Vishnuvardhan Road, Bengaluru – 560 098
DEPARTMENT OF ELECTRONICS & COMMUNICATION ENGINEERING
(Accredited by NBA for the academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
CERTIFICATE
This is to certify that the Project entitled “Bone cancer detection and classification using CNN” is
a bonafide work carried out by Manoj. P, bearing USN: 1RN19LVS02, in partial fulfillment for the
award of the degree of Master of Technology in VLSI DESIGN AND EMBEDDED SYSTEMS
affiliated to Visvesvaraya Technological University, Belagavi during the academic year 2020-
2021. It is certified that all corrections/suggestions indicated for internal assessment have been
incorporated in the report deposited in the department library. The project report has been approved
as it satisfies the academic requirements with respect to project work prescribed for the said degree.
..…………………. ..………………….. .....…………………
Dr. Manjula V K Dr. Vipula Singh Dr.M.K.Venkatesha
Internal guide HOD Principal
External Viva-Voce
Name of the Examiners Signature with Date
1……………………. …………………..
2…………………….. …………………..
RNS INSTITUTE OF TECHNOLOGY
(AICTE Approved, VTUAffiliated and NAAC ‘A’ Accredited)
(UG programs – CSE, ECE, ISE, EIE and EEE have been accredited by NBA
for the Academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
Channasandra, Dr. Vishnuvardhan Road, Bengaluru – 560 098
DEPARTMENT OF ELECTRONICS & COMMUNICATION ENGINEERING
(Accredited by NBA for the academic Years 2018 - 19, 2019 - 20 and 2020 - 21)
DECLARATION
I, Manoj P, bearing USN: 1RN19LVS02, hereby declare that the dissertation entitled “Bone
cancer detection and classification using CNN” which is being submitted by me as partial
fulfilment for the award of Master of Technology degree from Visvesvaraya Technological
University, Belagavi is an authentic record of my own work carried out during M.Tech final year at
RNSIT, Bengaluru under the supervision of my project guide Dr. Manjula V K, Associate
Professor, Department of Electronics & Communication Engineering
I further undertake that the matter embodied in the dissertation has not been submitted previously for
the award of any degree or diploma by me to any institution.
…………………………..
Date: 06/08/2021 Manoj P
Place: Bengaluru 1RN19LVS02
ACKNOWLEDGEMENT
The joy and satisfaction that accompany the successful completion of any task would be
incomplete without the mention of those who made it possible. I consider myself proud to be a part
of RNS Institute of Technology, the institution which stood by me in all my endeavors.
I express my gratitude to our beloved Chairman late Dr. R N Shetty and MD Satish R
Shetty for providing state of art facilities.
I would like to express our sincere thanks to Dr.M K Venkatesha, Principal and Dr.Vipula
Singh, Professor and Head, Department of ECE, for their valuable guidance and encouragement
throughout our program.
I would like to express my sincere gratitude to our guide Dr.Manjula V K, Associate
Professor for their guidance, continuous support and motivation in completing the project
successfully.
My profound gratitude to the P.G Coordinator Dr. Suresh D, Professor who has given
valuable suggestions and guidance throughout the project work.
I am also thankful to all the teaching and non-teaching faculty members of Department of
ECE, RNSIT for their constant support.
i
ABSTRACT
Cancer is a serious health problem among various kinds of diseases. More than one in
three people will be affected by some form of cancer during their lifetime. Among various types of
cancer, bone cancer is a leading cause of cancer-related death in many countries. In the U.S, 5-10%
of new cases of cancer are primary bone tumors. The most common type of primary malignant bone
tumor is osteosarcoma. The intention of the present work is to improve the detection and diagnosis
of bone cancer using computer-aided detection and diagnosis. A tool based approach such as
convolutional neural networks (CNN) can significantly decrease the surgeon’s workload and make
a better prognosis of patient conditions. CNNs need to be trained on a large amount of data in order
to achieve a more trustworthy performance.
In this study, learning techniques which are CNNs, are adapted to a public dataset 856
images on bone cancer histological images to detect cancerous, non-cancerous and healthy images.
First, the dataset will be preprocessed, and different classifications such as cancerous, non-
cancerous and healthy are applied. Then, model is trained on images and will be used for
classification. Experimental results show that training accuracy of 99% and validation accuracy of
84% and f1 score of 0.88 for non-viable image, 0.85 for healthy bone image and 0.79 for viable
bone images. Finally, the fine-tuned model demonstrates state-of-the-art performance and can be
used for detecting malignancy of Bone cancer based on histologic images.
ii
TABLE OF CONTENTS
Acknowledgement i
Abstract ii
List Of Contents iii
List Of Figures v
List Of Tables vi
List Of Acronyms vii
Publication viii
Chapter 1 : INTRODUCTION 1
1.1 Motivation 3
1.2 Necessity of System 3
1.3 Problem Statement 3
1.4 Objectives 4
1.5 Organization of Report 4
Chapter 2 : LITERATURE SURVEY 6
Chapter 3 : REQUIREMENTS 12
3.1 Non-Functional Requirements 12
3.2 Functional Requirements 13
3.3 Hardware and Software Requirements 13
Chapter 4 : METHODOLOGY 18
4.1 CNN Architecture 23
Chapter 5 : PROJECT DESCRIPTION 29
5.1 System Architecture 29
5.2 Data Flow Diagram 30
5.3 Use Case Diagram 31
Chapter 6 : TESTING 32
6.1 Testing Methods 32
6.2 Levels of Testing 32
Chapter 7 : RESULT 37
Chapter 8 : OBSERVATION AND ANALYSIS 40
iii
8.1 Confusion Matrix 40
8.2 Classification Report 41
Chapter 9 : CONCLUSION 44
9.1 Future Work 46
References 47
iv
LIST OF FIGURES
Figure 4.1 – Methodology 18
Figure 4.2 – Images used for classification 19
Figure 4.3 – Original Image 20
Figure 4.4 – Preprocessed Image 21
Figure 4.5 – CNN Architecture 24
Figure 4.6 – Convolution Layer 25
Figure 4.7 – Pooling Layer 26
Figure 4.8 – ReLU function 28
Figure 4.9 – Fully Connected Layer 28
Figure 5.1 – System Architecture 29
Figure 5.2 – Data Flow Diagram 30
Figure 5.3 – Use Case diagram 31
Figure 7.1 – Viable Tumor Input Image 37
Figure 7.2 – Viable tumor Preprocessed image with predicted class and accuracy of 37
prediction
Figure 7.3 – Non - Viable Tumor Input Image 38
Figure 7.4 – Non-Viable tumor Preprocessed image with predicted class and accuracy 38
of prediction
Figure 7.5 – Non - Tumor Input Image 38
Figure 7.6 – Non - tumor Preprocessed image with predicted class and accuracy of 39
prediction
Figure 8.1 – Confusion Matrix 41
v
LIST OF TABLES
Table 6.1 – User Test Case - I 34
Table 6.2 – User Test Case - II 34
Table 6.3 – Integration Test Case - I 35
Table 6.4 – Integration Test Case - II 36
Table 8.1 – Classification Report 42
Table 8.2 – Result Comparison 43
Table 9.1 – Output for Input Images 44
vi
LIST OF ACRONYMS
ACC Accuracy
ADF Anisotropic diffusion filter
ANN Artificial neural network
CNN Convolutional neural network
CT Computed Tomography
FN False Negative
FP False Positive
GDD Generalized Gaussian Density
IARC International Agency for Research on Cancer
MRI Magnetic Resonance Imaging
N Negative
NICPR National Institute of Cancer Prevention and Research
P Positive
ReLU Rectified Linear Unit
RNN Recurrent neural network
RSS – BW W bidirectional network, which is semi‐supervised regenerative
SVM-RBF Support vector machine –radial basis function
TN True Negative
TP True Positive
VGG Visual geometry group
vii
Publication
The project carried out has been resulted in two publications as mentioned below.
Manoj P and Dr Manjula V K, “Bone cancer detection using Convolutional Neural Networks”,
IETE Sponsored Fourth National Conference on Emerging Trends in Engineering science and
Technology, 4th NCETEST July 2021, pp 5012 - 5016 .
Manoj P and Dr Manjula V K, “Bone Cancer Detection Using Convolution Neural Network –
An overview”, IJCRT, March 2021.
viii
