Professional Documents
Culture Documents
Artigo 10.1109@ICCSDET.2018.8821067
Artigo 10.1109@ICCSDET.2018.8821067
Abstract—Leukaemia is an abnormal proliferation of White [5]. These symptoms are due to reduced count of normal blood
Blood Cells in the bone marrow and peripheral blood. The cells [1].
diagnosis of leukaemia is done by the pathologists and
haematologists by observing a peripheral blood smear and bone On the basis of cell-line of origin, Leukaemia is classified
marrow smear under a microscope. Advanced cytochemistry into Myeloid or Lymphoid and based on the clinical course it
tests, immunophenotyping, flow cytometry, chromosome analysis is classified into Acute or Chronic. This leads to four main
etc. are used to confirm the incidence of leukaemia. Since these categories of leukaemia; Acute Myeloid Leukaemia (AML),
tests are time intense and exorbitant, image analysis of the blood Chronic Myeloid Leukaemia (CML), Acute Lymphoblastic
and bone marrow smear can be used to assist the pathologists Leukaemia (ALL) and Chronic Lymphocytic Leukaemia
and haematologists for a quick diagnosis of leukaemia. A simple (CLL). AML is the most common form of acute leukaemia
method to detect the incidence of leukaemia automatically by during the first few months of life. It is rare in childhood and
image processing techniques is proposed in this paper. The adolescence but it is the most frequently observed acute
method uses colour based image segmentation using k-means leukaemia in middle and later years of life. CML occurs
clustering and image analysis by Hausdorff Dimension to detect primarily in young and middle aged adults. ALL occurs
leukaemia. The image in the RGB colour model is converted to primarily in children between two and ten years of age. It also
the CIELAB as well as CMYK colour spaces and a comparison occurs in middle-aged and elderly adults. Most cases of CLL
between the two in detecting leukaemia is also presented.
occurs over the age of 60 and it is more than twice as common
Keywords—Leukaemia, segmentation, Hausdorff in men as in women [1].
Dimension, k- means, clustering The diagnosis of Leukaemia is usually done by the
haematologists and pathologists by manually observing a
I. INTRODUCTION blood smear of the patient under a microscope. An abnormal
WBC count may be an indication of leukaemia [1, 2]. The
Leukaemia is a type of cancer that affects blood cells and count and structure of the blast cells compared to the normal
sometimes commonly known as Blood Cancer. Leukaemia cells are used to identify and classify leukaemia. The
usually originates in the bone marrow and leads to an leukaemia can be confirmed and the specific type can be
abnormal spread of White Blood Cells (WBCs). These WBCs identified by a needle biopsy and aspiration of bone marrow
are not completely developed and are known as blasts [1]. The from a pelvic bone. The conventional method is time
exact cause of Leukaemia is not known to the medical world consuming and greatly influenced by the skill, ability and
and may be due to inherited factors and environmental factors. fatigue level of the persons involved in the diagnostic
About 62130 cases of leukaemia is diagnosed in the U.S. in procedures. Advanced methods such as cytochemistry tests,
2017 [33]. There are around 9900 new Leukaemia cases in the immunophenotyping, chromosome tests etc. can also be done
U.K. every year [34]. In India, more than 10,000 cases of to confirm leukaemia but are expensive and time consuming
childhood Leukaemia are reported every year. for a routine examination [6, 7, 8].
Image processing and analysis of the peripheral blood
The human blood comprises of three types of cells; smear and bone marrow smear images can be used to assist the
Erythrocytes or Red Blood cells (RBCs), Leukocytes or White haematologists and pathologists to identify many diseases like
Blood Cells (WBCs), and Thrombocytes or Platelets. About malaria, skin diseases, leukaemia etc. Automated counting of
45% of whole blood is RBC, about 54.3% is plasma and about blood cells can also be implemented by image processing
0.7% is WBC. Leukaemia results in large number of abnormal techniques. These methods are fast, easy and less expensive
WBCs and this in turn leads to reduced count of normal blood and help the pathologists for a quick diagnosis. A remote
cells. Symptoms of Leukaemia are bleeding, bruising diagnosis is also very easy as the image can be transmitted to
problems, fatigue, fever and an increased threat of infections any location over the internet rather than sending the blood
sample.
Images
Hausdorff Dimension VI. ACKNOWLEDGMENT
(Leukaemia cases) L*a*b* CMYK We would like to express our sincere gratitude to the
American Society for Hematology for their online data bank
Sample 1 1.3311 1.3271
of high quality image samples.
Sample2 1.3189 1.3743
Sample 3 1.3756 1.3868 REFERENCES
Sample 4 0.9951 1.0048
Sample 5 1.6024 1.6231 [1] John Bernard Henry, Clinical Diagnosis and Management by Laboratory
Methods W.B. Saunders Company, 17th Edition, Philadelphia, 1989.
Sample 6 1.4075 1.3956 [2] Sos Agaian, Monica Madhukar and Anthony T. Chronopoulos,
Sample 7 1.2042 1.183 “Automated screening system for Acute Mylogenous Leukaemia
detection in Blood microscopic images,” IEEE SYSTEMS JOURNAL,
Sample 8 0.9980 0.9999 Vol. 8, No.3, pp. 995-1004, September 2014.
Sample 9 1.3641 1.3577 [3] Fabio Scotti, “Automatic morphological analysis for Acute Leukaemia
identification in Peripheral Blood microscopic images,” in Proceedings
Sample 10 1.6525 1.6096 of IEEE International Conference on Computational Intelligence for
Measurement Systems and Applications, pp. 96-101, 2005.
Sample 11 1.3282 1.2984
[4] Vincenzo Piuri, Fabio Scotti, “Morphological classification of Blood
Sample 12 1.2651 1.2422 Leucocytes by microscope images,” in Proceedings of IEEE
International Conference on Computational Intelligence for
Sample 13 1.2042 1.183 Measurement Systems and Applications, pp.103-108, 2004.
Sample 14 0.9974 0.9968 [5] Subrajeet Mohapatra, Dipti Patra and Sanghamitra Satpathy,
“Unsupervised blood microscopic image segmentation and Leukaemia
Sample 15 1.1625 1.1478 detection using colour based Clustering,” International Journal of
Computer Information Systems and Industrial management
Sample 16 1.2907 1.2816 Applications, Vol.4 pp. 477-485, 2012.
Sample 17 1.3825 1.3204 [6] Subrajeet Mohapatra and Dipti Patra, “Automated cell nucleus
segmentation and Acute Leukaemia detection in Blood microscopic
Sample 18 1.5423 1.5049 images,” in Proceedings of International Conference on Systems in
Medicine and Biology,pp. 49-54, 2010.
Sample 19 1.3321 1.3449
[7] Subrajeet Mohapatra, Sushanta Shekhar, Dipti Patra and Sanghamitra
Sample 20 1.4253 1.431 Satpathy, “ Fuzzy based Blood image segmentation for automated
Leukaemia detection,” in Proceedings of ICDeCom, pp. 1-5, 2011.
Mean 1.30895 1.300645 [8] Ashwini Rejintal and Aswini N, “Image processing based Leukaemia
cancer cell detection,” in Proceedings of IEEE International Conference
On Recent Trends In Electronics Information Communication
The mean of the HD of the 20 cases of leukaemia samples in Technology, pp. 471-474, 2016.
CIELAB colour model is 1.30895 and that of the CMYK [9] Farnoosh Sadeghian, Zainia Seman, Abdul Rahman Ramli, Badrul
colour space is 1.300645. That is, the mean HD value of the Hisham Abdul Khader and M-Iqbal Saripan, “A framework for White
leukaemia cases is almost same in CIELAB and CMYK Blood Cell segmentation in microscopic Blood images using Digital
Image Processing,” Biological Procedures Online, Vol. 11, No. 1, 2009.
colour models. This indicates that processing the image in
[10] Nurul Haswani Abd Halim, Mohd Yusoff Mashor, and Rosline Hassan,
CIELAB and CMYK colour spaces is almost equally good in “Automatic Blast counting for Acute Leuklemia based on Blood
isolating the leukaemia cases using Hausdorff Dimension. Samples,” International Journal of Research and Reviews in Computer
Science, Vol. 2, No. 4, pp. 971-976, 2011.
V. CONCLUSION [11] Shaikh Mohammed Bilal N and Sachin Deshpande, “Computer aided
The paper reported the use of Hausdorff Dimension Leukaemia detection using Digital Image Processing techniques,” in
Proceedings of IEEE International Conference On Recent Trends In
(HD) in automated detection of leukaemia by analysing the Electronics Information Communication Technology, pp. 344-348,
blood smear images and compares processing of the images in 2017.
CIELAB and CMYK colour spaces. The work used 25 good [12] M.A. Khosrosereshki and M.B. Menhaj, “A fuzzy based classifier for
quality images obtained from the image bank of American diagnosis of acute lymphoblastic leukaemia using blood smear image
processing,” in Proceedings of 5th Iranian Joint Congress Fuzzy and
Society of Hematology. The system automatically detects Intelligent Sytems(CFIS), pp. 13-18, 2017.
leukaemia by comparing the HD of the segmented image [13] A.R. Jasmine Begum and T. Abdul Razak, “Diagnosing Leukaemia from
sample with the mean HD of the normal image samples. It microscopic images using image analysis and processing techniques,” in
successfully detected the incidence of leukaemia in all the 20 Proceedings of World Congress on Computing and Communication
cases considered, by comparing the HD with the mean of the Technologies, pp. 227- 230, 2017.
normal cases. The study also found that processing and [14] Vasuki Shankar, Murali Mohan Deshpande, N Chaitra, and S Aditi,
“Automatic detection of Acute Lymphoblastic Leukaemia using Image
analyzing the images in CIELAB colour model is Processing,” in Proceedings of IEEE International Conference on
comparatively better in producing good quality segmented Advances in Computer Applications, pp. 186-189, 2016.
images. [15] Biji G. and S. Hariharan, “An efficient Peripheral Blood Smear image
analysis technique for Leukaemia detection,” in Proceedings of
International Conference on IoT in Social, Mobile, Analytics and Cloud, [23] Sasirassamee Buavirat and Chonawat Srisa-an, “Classification for Acute
pp. 259- 264, 2017. Lymphocytic Leukaemia using feature extraction and Neural Networks
[16] Hayan T. Madhloom, Sameem Abdul Kareem, and hany Ariffin, “A in White Blood Cell stained images,” in Proceedings of The 3rd
robust feature extraction and selection method for the recognition of International Symposium on Biomedical Engineering, pp. 1-4, 2008.
Lyphocytes versus Acute Lymphoblastic Leukaemia,” in Proceedings of [24] Firdaus Ismail Sholeh, “White Blood Cell segmentation for fresh Blood
International Conference on Advanced Computer Science Applications smear images,” in Proceedings of ICACSIS, pp. 425-429, 2013.
and Technologies, pp. 330-335, 2012. [25] Dorin Comaniciu and Peter Meer, “Advanced Algorithmic Approaches
[17] S. Ravikumar and A. Shanmugam, “WBC image segmentation using to Medical Image Segmentation,” pp. 541-558, Springer, 2001.
modified Fuzzy Possibilistic C-Means algorithm,” Journal of [26] Viswanathan P, “Fuzzy C means detection of Leukaemia based on
Theorectical and Applied Information Technology, Vol.63, No.2, 2014. Morphological Contour Segmentation,” in proceedings of Second
[18] Stefan Skrobanski, Stelios Pavlids, Waidah Ismail, Rosline Hassan, Stev International Symposium on Computer Vision and the Internet, 2015.
Counsell, and Stephen swift, “Use of general purpose GPU [27] B.T. Milne, “Measuring the fractal geometry of landscapes,” Applied
programming to enhance the classification of Leukaemia Blast cells in Mathematics and Computation, Vol. 27, 1988.
Blood smear images,” in Proceedings of IDA 2012, pp. 369-380.
[28] A.P. Pentland,“Fractal based description of natural scenes,” IEEE
[19] Reymond Joseph A. Cabrera, Criselle Amor P. Legaspi, Erika Jasmine transactions on Pattern Analysis and Machine Intelligence, (6),1986.
G. Papa, Reden D. Samonte, and Donata D. Acula, “HeMatic: An
automated Leukaemia detector with separation of overlapping Blood [29] ASH Image Bank: American Society of Hematology,
http://imagebank.hematology.org.
cells through Image Processing and Genetic Algorithm,” in Proceedings
of IEEE- ICAST, pp. 985-987, 2017. [30] Rafael C. Gonzalez and Richard E. Woods, Digital Image Processing,
Pearson Education Asia, Second Edition, 2002.
[20] Preeti Jagadev and H.G. Virani, “Detection of Leukaemia and its types
using Image Processing and Machine Learning,” in Proceedings of [31] A.K. Jain, Fundamentals of Digital Image Processing, Pearson
International Conference on Trends in Electronics and Informatics, pp. Education, First Edition, 2003.
522-526, 2017. [32] R Panneerselvam, Research Methodology, PHI, Second Edition, 2014.
[21] Lim Huey Nee, Mohd Yusoff Mashor, and Rosline Hassan, “White [33] The Leukaemia & Lymphoma Society, New York, https:// www.ils.org
Blood cell segmentation for Acute Leukaemia bone marrow images,” in [34] Cancer Research UK, http:// www. cancerresearchuk.org.
Proceedings of International Conference on Biomedical Enginnering, pp.
357-361,2012.
[22] Fauziah Kasmin, Anton Satriya Prabuwono, and Azizi Abdullah,
“Detection of Leukaemia in human Blood sample based on microscopic
images: A study,” Journal of Theoretical and Applied Information
Technology, Vol. 46, No. 2, pp.579-586, 2012.