Detection and Classification and Infarct and Hemorrhage for

Stroke

Myat Mon Kyaw

University of Technology, Yatanarpon Cyber City

A thesis submitted to the University of Technology, Yatanarpon Cyber City

in fulfillment of the requirement for the degree of
Doctor of Philosophy

2014
 Statement of Originality

I hereby certify that the work embodied in this thesis is the result of

original research and has not been submitted for a higher degree to any

other University or Institution.

Date Myat Mon Kyaw

 Acknowledgments

First of all, I would like to express my deep gratitude to Dr. Ko Ko Oo, Union
Minister of the Ministry of Science and Technology, for opening of Special Courses for
Doctoral Program at University of Technology (Yatanarpon Cyber City).
I would like to thank to Dr. Aung Win, Principal and Professor, University of
Technology (Yatanarpon Cyber City) who gave me an opportunity to work on this project
which allows me to gain much research experience.
Furthermore, my deepest thanks and appreciation go to [Link] Soe Khaing, Professor
and Head of Information and Communication Technology, and my respectful teachers who
specifically offered strong morale and physical support, care and kindness during the years of
my studies.
I would like to thank to Dr. Su Su Maung, Associate Professor, for her kindly
guidance, encouragement and supervision to do this research. I would like to express special
appreciation and heartfelt thanks to my supervisor for her effective advices and helpful
suggestions.
My gratitude goes to medical radiologists, Dr. Zarni Lwin, Head of Radiology
Department, Pyin Oo Lwin Civil Hospital and Dr. Lin Lin Htun, Head of Radiology
Department, Yangon General Hospital, for their support for acquiring relevant brain CT
images, related information and the valuable diagnosis decisions.
Finally, thanks are also extended especially to my parents, colleagues, friends and staff
of University of Technology (Yatanarpon Cyber City), for their contribution to the successful
completion of this thesis.

I
 Abstract

Detection and Classification and Infarct and Hemorrhage for Stroke

The proposed system is automatic method for the detection and classification of an
abnormality into infarct and hemorrhagic or normal in brain CT images. A stroke occurs when
a blood vessel either bursts or there is a blockage of the blood vessel. Stroke results in serious
long term disability or death as an affected area of the brain cannot function. Strokes are
mainly classified in two categories: infarct stroke and hemorrhagic stroke. Firstly, the original
image is converted into gray scale and noise and film artifacts are removed by using a median
filter. Then skull regions are removed by a morphological function. Because skull portions are
not required for further processing and have similar pixel distribution with haemorrhage
region. In classification, the important aspect is to select the best features for the identification
of abnormality criteria. Image is classified into infarct or hemorrhagic stroke or normal image
based on features from gray-level co-occurrence matrix, total number of pixels, mean and
standard deviation of the image. The abnormality contained hemisphere is also detected based
on these features. The accuracy of the classification results can be compared with the
specialists’ decision or radiologist labelled ground truth. Segmentation method is applied to
determine the accurate region of abnormality. Histogram-based thresholding method is
applied to extract region of interest (i.e., abnormality) from parts of the brain by combining
with contour. In histogram-based thresholding, the threshold is derived from the histogram.
The seeded region growing method is also applied for abnormal region detection. The initial
seed is derived from the threshold in histogram-based thresholding. In the experiment, 95 CT
images are used. 67 images are hemorrhagic stroke affected image, 16 are infarct stroke
affected image and others are abnormal image. These images are classified with accuracy
98.95%. The parts of the brain are segmented and an abnormal region is extracted.

This work is done in addition to a Software Development Project. Both of these are carried out in partial
fulfillment of the requirements for the Degree of Doctor of Philosophy in University of Technology, Yatanarpon
Cyber City.
ii
 Table of Contents

Acknowledgments i

Abstract ii

(Actual Table of Content to be placed here – this page) .

List of Figures .

List of Tables .

List of Equations (if any) .

1. Introduction (include the following sub-headings) 1

1.1 Motivation ………………………………………………. 1

1.2 Objectives ………………………………………………. .
1.3 Major Contribution of Thesis .
1.4 Organization of the Dissertation ………………………… .

2.

. Actual work done and contribution, including literature survey.

. (It will normally contain 4 to 5 chapters)
.

N. Conclusion and Recommendations

N.1 Conclusion
N.2 Recommendations for further research

Author's Publications

Bibliography

Appendices (if any)

Author's Publications (sample format)

[1] N. Q. Hung, S. Y. Lee, L.X. Hung, "A Middleware Framework for Context Acquisition
in Ubiquitous computing Systems", Proceedings of 2nd International Conference on
Computer Application 04, January 8, 2004, Myanmar, pp. 209-216.

[2] Soe Soe Khaing and Thinn Thu Naing "A Real-Time Packet Scheduling Model for
Multimedia Applications in Packet-Based Network” In Proceeding of the Fifth
International Conference on Computer Application, Myanmar, February 8-9, 2007, pp-
51-60.

[3] Soe Soe Khaing, "Green in ICT Utilization towards Sustainable Development of
Myanmar", Proceedings of 2nd Regional Conference on Informaton and Communicatoin
Technology, Maarch 2-3, 2010, Indonesia, pp. 151-158.
Bibliography (sample format)

[1] W S Kim and A K Bejczy, “Demonstration of a High-Fidelity Predictive/Preview

Display Technique for Telerobotic Servicing in Space”, IEEE Trans. On Robotics and
Automation, vol.9, no.5, 1993, pp. 698-702.

[2] G V Paul and K Ikeuchi, “Quasi-linear method for computing and projecting onto C-
surfaces: Planar case”, Proceedings of IEEE International Conference on Robotics and
Automation, USA, 1997, pp. 2032-2037.
 UNIVERSITY OF TECHNOLOGY, YATANARPON CYBER CITY
Format of the PhD Thesis

- Page limit : ( minimum 100 pages )

- International Standard Paper Size A4 (297 x 210mm) should be used.
- The typing should be (1.5) spaced. Use the Font "Time New Roman", Size 12.
May be presented on one side of the paper.
- Right, Top and Bottom margins should be 1 inch respectively : Left margin should
be 1.5 inch.
- Folding diagrams or charts should be arranged so as to open to the top and right.
- Chapter number, chapter title should be Time New Roman, Size-14, Bold face,
centered, 1.5 line spacing and must always begin on a new page. One line spaced
should use between the chapter title and main text.
- Paragraph indents should be 0.5 inch.
- Sub-titles of each chapter should be Time New Roman Font, Size 12, and Bold
faced.
- Sub-title and its first line of the paragraph should be of 6 point.

Order of contents:
- The thesis must be preceded by a title page.
- Statement of originality (Declaration)
- Acknowledgements
- Abstract
- Table of contents
- List of Figures
- List of Tables
- List of Equations (if any)
- Main Text
- Author's publications
- Bibliography
- Appendixes
An abstract should be of 300 - 700 words. It should summarize the appropriate headings,
aims, scope and conclusion of the thesis.
CLASSIFICATION AND DETECTION OF INFARCT AND
HEMORRHAGE FOR STROKE

MYAT MON KYAW

The University of Technology, Yatanarpon Cyber City

A thesis submitted to the University of Technology, Yatanarpon Cyber City

in fulfillment of the requirement for the degree of
Doctor of Philosophy

August, 2014
 Before producing final copies of the thesis for submission, the candidate should ensure
that all the spelling, grammar, punctuation and choice of language are of a doctoral standard
and the bibliography is complete and exact.

Tables and Figures

The characters and numbers in tables graphs and charts should be : Font- "Time New
Roman" and Size-10. Each table and figure must have a unique number assigned to it, along
with a title or description of the table or figures; Font Size 10 and Bold faced. Use the
numbering scheme by chapter wise, e.g., Table 1.1, 1.2,2.1, 2.2 and etc.

Pagination

Preliminary pages: Copy right page (if included) is not numbered. The declaration,
acknowledgement, preface, table of contents, and lists of tables and/or figures are numbered
consecutively in lower case Roman numeral at the bottom center of each page; e.g ( i , ii etc.)

Text: Use Arabic number beginning with '1' on the first page of the text and continuing
throughout the manuscript. The page number should be at the bottom center of the page
throughout the text.

Line Spacing of the text : The basic text of all dissertation must be 1.5.

Footnotes and references: Footnote indicators must be superscript. Footnote numbering

must be continuous within each chapter or appendix, not throughout the text. Must insert an
extra line space between notes when there is more than one on a page. A short line rule called
a separator dividing the footnote area from the main text is required.
Reference Format
- Sorted by First Author's Family name, ascending orders
e.g J. T. L. Wang
Family Name
- font size – 12
- line spacing – 1.5
- paragraph spacing - After-12 pt.

1. Author Name

Author Name should be of the form..

Kam Shu Kai Andy  K. S. K. Andy

2. Book

[xx] Author name1, Author name2 and Author name3, Book Title, Publisher/Press,
Year.

Example:
[1] N. Carriero and D. Gelernter. How to Write ParallelPrograms: A First Course. MIT
Press, 1990.

3. Transaction / Journal Paper

[xx] Author name1, Author name2 and Author name3, “Paper Title”, IEEE
Transaction ………., Vol. xx, No. xx, pp. xx-xx, Month, Year.

Example:
[1] S. Basagni, D. Bruschi, and I. Chlamtac. “A mobility transparent deterministic
broadcast mechanism for ad hoc networks”, IEEE/ACM Transactions on
Networking, vol. 7, no. 6, pp. 799–807, December 1999.

4. Conference/ Proceeding Paper

Author name1, Author name2 and Author name3, “Paper Title”, in Proceedings of
xxxxxxxx, Month, Year. pp. xx-xx .

Example:
[11] J. T.-L. Wang, G.-W. Chirn, T. G. Marr, B. Shapiro, D. Shasha, and K. Zhang,
“Combinatorial Pattern Discovery for Scientific Data: Some Preliminary
Results”, in Proc. of the 1994 ACM SIGMOD Intl. Conf. on Management of
Data, May 1994. pp. 102-115.

[12] G V Paul and K Ikeuchi, “Quasi-linear method for computing and projecting
onto C-surfaces: Planar case”, Proceedings of IEEE International Conference on
Robotics and Automation, USA, 1997, pp. 2032-2037.