978-1-5386-7942-5/19/$31.00 ©2019 IEEE
Automated Cell Counting System For Chronic
Leukemia
Nor Hazlyna Harun
School of Computing
Universiti Utara Malaysia
Sintok, Kedah, Malaysia Country
hazlyna@uum.edu.my
Nur Azzah Abu Bakar
School of Computing
Universiti Utara Malaysia
Sintok, Kedah, Malaysia
nurazzah@uum.edu.my
Uvaghesvary A/P Mohan
School of Computing
Universiti Utara Malaysia
Sintok, Kedah, Malaysia Country
uvagesmohan95@gmail.com
Maslinda Mohd Nadzir
School of Computing
Universiti Utara Malaysia
Sintok, Kedah, Malaysia
maslinda@uum.edu.my
Mohamad Ghozali Hassan
School of Technology Management and
Logistic
Universiti Utara Malaysia
Sintok, Kedah, Malaysia
ghozali@uum.edu.my
Robiyanti Adollah
Mechatronic School of Engineering
Universiti Malaysia Perlis
Arau, Perlis, Malaysia
robiyanti@unimap.edu.my
Abstract— Leukemia is a group of cancers which create a
large amount of immature white blood cells. Abnormal
numbers of white blood cells may suggest a screening of
leukemia, and the blood sample is examined under the
microscope to observe if the cells appear abnormal. The
manual screening of chronic leukemia is time consuming and
tedious while the Automated Hematology Analyzer is too
expensive, particularly for the third world countries. This has
been made exacerbated by the gold standard of biopsy
inspiration which is painful and invasive for the patient. An
automated cell counting (ACC) system for chronic leukemia
has been developed to support and ease the routine of
hematologist and technologist in the screening process and to
give a quick and accurate result. The fusion of image
processing technique has been proposed, which include four
main stages, i.e. image acquisition, image segmentation, noise
removal and counting process. Based on the sensitivity test
over 100 images of chronic cells, an overall result shows
98.94% sensitivity of the system performance and the
processing time recorded is less than 6 second per image. This
proved an excellent level of ACC system performance. It is
concluded that the system is suitable to be used as an
automated counting system for chronic leukemia disease due to
its sensitivity and ability to reduce the time taken for screening
process.
Keywords—Automated Cell Counting System (ACC),
Segmentation, Morphological Operation, Blood Cell.
I. INTRODUCTION
Leukemia is a group of cancers which create a large
amount of immature blood cells. These immature blood cells
take up space in the bone marrow, preventing the bone
marrow from making healthy blood cells such as platelets,
red blood cells, and white blood cells. Leukemia can be
classified into two, which are chronic leukemia and acute
leukemia [1]. Acute leukemia grows rapidly and will spread
over the body within a short period. Whereas, chronic
leukemia grows slower than acute leukemia and become
more serious over years [2].
According to [3], screening process based on microscopic
blood images appeared to have an error rate between 30 to
40% depending on the haematologists’ experience and also
the difficulties to distinguish between the normal and
abnormal cells [4].
The utilization of image processing technique as an
automated counting system for diagnosis of leukemia disease
has been studied in [5] and [6] in which the image
segmentation technique such as K-mean clustering is
employed in several leukemia cases. In [7], K-mean
clustering is used for automated acute lymphocytic leukemia
detection. While in [8], K-mean clustering is integrated with
histogram equalization and Zack algorithm to determine
leukemia. Recent study also used K-mean clustering to
identify the leukemia cells and features extraction, and image
renement [9]. Another study presented the integration of K-
mean clustering and machine learning technique to
differentiate among Acute lymphoblastic leukemia (ALL),
Acute myeloid leukemia (AML), Chronic lymphocytic
leukemia (CLL) and Chronic myeloid leukemia (CML) [10].
What is learnt from the above is the effectiveness of
image processing technique in automating leukemia
diagnosis. However, most of the studies in the field of image
processing only focused on cases outside Malaysia. Hence,
this study serves as a replication study which focuses on
leukemia cases in Malaysia. The key contribution of this
study is the development of an ACC for Chronic Leukemia
to improve the reliability of the counting blood cell and
decrease the dependency on human experts.
ACC is introduced to automatically count chronic
leukemia cell and at the same time promised an acceptable
result for screening process based on blood sample. Details
of the development of the system is discussed in Section II
and III. Section IV of this paper concludes the work that is
presented in the earlier sections.
II. DEVELOPMENT OF ACC SYSTEM
ACC was developed using MATLAB environment which
utilized image processing technique for chronic cell counting
based on the blood samples. There are four stages including
image acquisition, image segmentation, noise removal and
counting process. Fig. 1 indicates the combination of image
502
brought to you by CORE View metadata, citation and similar papers at core.ac.uk
provided by UUM Repository