By
Yulius Wibisono 11502005
BACHELOR’S DEGREE in
INFORMATION TECHNOLOGY
ENGINEERING AND INFORMATION TECHNOLOGY
SWISS GERMAN UNIVERSITY The Prominence Tower
Jalan Jalur Sutera Barat No. 15, Alam Sutera Tangerang, Banten 15143 - Indonesia
Revision after Thesis Defense on 11 July 2019
Yulius Wibisono
STATEMENT BY THE AUTHOR
I hereby declare that this submission is my own work and to the best of my knowledge, it contains no material previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at any educational institution, except where due acknowledgement is made in the thesis.
Yulius Wibisono
_____________________________________________
Student Date
Approved by:
Anto Satriyo Nugroho, Dr.Eng.
_____________________________________________
Thesis Advisor Date
Dr. Maulahikmah Galinium, S.Kom., M.Sc.
_____________________________________________
Thesis Co-Advisor Date
Dr. Maulahikmah Galinium, S.Kom., M.Sc.
_____________________________________________
Dean Date
Yulius Wibisono
ABSTRACT
OPTIMIZATION ON MALARIA COMPUTER AIDED DIAGNOSTIC SYSTEM By
Yulius Wibisono
Anto Satriyo Nugroho, Dr.Eng., Advisor
Dr. Maulahikmah Galinium, S.Kom., M.Sc., Co-Advisor SWISS GERMAN UNIVERSITY
In the field, malaria infection are analyzed by manually examining a thin blood smear that is acquired from the infected patient. This method requires a trained human interaction and therefore it is time consuming and prone to errors. A development of Computer Aided Diagnostics (CAD) for Malaria were conducted before. The result of the program is able to recognize the infected red blood cells, the species and the life phase of the infecting Plasmodium. However, the average runtime of the program is 41.45 seconds per image, which is too long if the program is to be used in the field. By measuring the runtime of each process in the program, optimization can be done by re- writing or substituting the algorithm that causes the longest runtime. The optimization to the Connected Component Labelling process by substituting it with Contour Tracing Labelling caused the program to be 8 times faster. By reducing the resolution of the image before the Clump Splitting process results in speedup by 14 times faster.
Applying Concave Point Based Clump Splitting speeds up the runtime by 23 times faster compared to the original CAD. Optimization is successfully performed so that the CAD has an average runtime of 1.73 seconds while only affecting accuracy by a small margin, from 63% to 74% on the Infected vs Healthy Classification, 66% to 71%
on the Species Classification, and from 73% to 67% on the Life Phase Classification.
Keywords: malaria, thin blood smear, contour tracing labelling, clump splitting.
Yulius Wibisono
© Copyright 2019 by Yulius Wibisono
All rights reserved
Yulius Wibisono DEDICATION
To God, for His blessing.
For my parents, for their ever-loving support.
For my friends, who believes in me.
Yulius Wibisono
ACKNOWLEDGEMENTS
I would like to thank God for His blessing until this point of time that I could write this research.
I also would like to thank my parents, who has been really supportive of what I’ve done and for their unconditional love to me. I hope I can return what they have given to me someday.
This research would not have been completed if it were not for Anto Satriyo Nugroho, Dr.Eng., my Advisor’s effort on this malaria research since 2011. His assistance throughout my struggle in doing this research and his patience really helped me to keep moving forward and discover things that at first I thought I was not capable of. I also express my huge gratitude to Dr. Maulahikmah Galinium, S.Kom., M.Sc., my Co- Advisor in this research for his assistance and feedbacks in writing this research. I also thank Tommy Winarta for his encouragement to continue this study.
The dataset provided by Eijkman Institute of Molecular Biology really helps and is one of the most important requirement in this research, and for that I would like to express my gratitude.
The amazing lecturers of IT class has really helped me surviving the rough 4 years learning in SGU. I really thank to those who has patiently shared their knowledge and has given their very inspiring insights.
I also would like to thank my friends, because they really helped to make my experience throughout my studies in SGU very colorful and memorable because of the jokes we crack together, phones/bag/drinking bottle/ID Card that we hid together, and for the weight that we pull together.
Yulius Wibisono
TABLE OF CONTENTS
Page
STATEMENT BY THE AUTHOR ... 2
ABSTRACT ... 3
DEDICATION ... 5
ACKNOWLEDGEMENTS ... 6
TABLE OF CONTENTS ... 7
LIST OF FIGURES ... 10
LIST OF TABLES ... 13
CHAPTER 1 - INTRODUCTION ... 15
1.1. Background ... 15
1.2. Research Problems ... 16
1.3. Research Objectives ... 16
1.4. Significance of Study ... 17
1.5. Research Questions ... 17
1.6. Hypothesis... 17
CHAPTER 2 – LITERATURE REVIEW ... 18
2.1. Malaria ... 18
2.2. Thin Blood Smear and Image Acquisition... 18
2.3. Related Works ... 20
2.3.1. Red Blood Cell Segmentation (Anggraini et al., 2011)... 21
2.3.2. Clumped Cell Splitting Algorithm (Setyono, 2015) ... 22
Yulius Wibisono
2.3.3. Malaria Plasmodia Status Identification (Anggraini et al., 2011) ... 23
2.3.4. Plasmodium Species Classification (Chandra et al., 2017) ... 23
2.3.5. Species and Phase Classification of Plasmodium (Winarta, 2017) ... 24
2.4. Algorithms ... 26
2.4.1. Connected Component Labelling ... 26
2.4.2. Contour Tracing Labelling (Chang, Chen, & Lu, 2004) ... 26
2.4.3. Hole Filling ... 28
2.4.4 Clump Splitting ... 32
CHAPTER 3 – RESEARCH METHODOLOGY ... 34
3.1. Optimization ... 34
3.1.1. Thin Blood Smear Microphotograph Preparation ... 35
3.1.1. Testing the Existing Algorithm with New Environment to Get New Benchmarks ... 35
3.1.2. Algorithm Analysis ... 36
3.1.3. Prototype Code Development... 40
3.1.4. Testing Against Existing Algorithm ... 40
3.1.5 Proposed Algorithms ... 41
CHAPTER 4 – RESULT AND DISCUSSION ... 61
4.1. Existing CAD (Program A) ... 62
4.2. Code Optimization Result ... 65
4.2.1. Optimization using Connected Component Labelling with Bounding Box (Program B) ... 65
Yulius Wibisono 4.2.2. Optimization Using Contour Tracing Labelling as Connected Component
Labelling Substitute (Program C) ... 69
4.2.3. Optimization Using Downscaled Clump Splitting as Clump Splitting Substitute (Program D) ... 72
4.2.4. Optimization by Using Downscaled Concave Point Based Clump Splitting (Program E) ... 75
4.2.5. Overall Optimization Result ... 80
4.3 Accuracy Analysis ... 84
4.3.1. Healthy vs Infected Classification ... 85
4.3.2. Species Classification ... 92
4.3.3. Phase Classification ... 96
4.4. Error Analysis ... 100
4.5. Graphical User Interface ... 100
CHAPTER 5 – CONCLUSION AND RECOMMENDATIONS ... 107
5.1. Conclusions ... 107
5.2. Recommendations and Future Works ... 107
GLOSSARY ... 109
APPENDICES ... 112
Appendix 1: Species Classification Data of the Optimized CAD without any omitted data ... 112
Appendix 2: Phase Classification Data of the Optimized CAD without any omitted data ... 114
Appendix 3: Pseudocode for Concave Point Based Clump Splitting ... 116
CURRICULUM VITAE ... 118