See discussions, stats, and author profiles for this publication at: [Link]

net/publication/230815213

An Accurate and Cost Effective Approach to Blood Cell Count

Article  in  International Journal of Computer Applications · August 2012

DOI: 10.5120/7734-0682

CITATIONS READS

25 5,100

2 authors, including:

Aamir Khan
University of Glasgow
14 PUBLICATIONS   97 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Human Activities Monitoring using Deep Learning View project

All content following this page was uploaded by Aamir Khan on 17 January 2017.

The user has requested enhancement of the downloaded file.

 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

An Accurate and Cost Effective Approach to Blood Cell

Count
Sanaullah Khan Aamir Khan Faisal Saleh Khattak,
Department of Electrical Department of Electrical Arslan Naseem
Engineering, Comsats Institute Engineering, Comsats Institute Department of Electrical
of Information Technology Wah of Information Technology Wah Engineering, Comsats Institute
Campus, WahCantt. Pakistan. Campus, WahCantt. Pakistan. of Information Technology Wah
Campus, WahCantt. Pakistan.

ABSTRACT one cant neglect the chance of error in the report due to error
Medical imaging brings up the methods and technique which caused by apparatus, personal errors, statistical errors etc.
enable us to take and analyse images of different disease and While on the other hand latest haematology analyzer
human body parts for diagnosing purposes. The old somehow error free and fast but it is widely unavailable and
conventional methods used in hospital laboratories to detect very expensive machine and the countries like Pakistan are
and diagnose the problem were very slow and they had greater resource less to provide it in every hospital laboratory in
chances of error in it but medical imaging through different country. So as a result of the problem this research based
image processing techniques opened new, fast and cost- project proposed a new method of cell counting which is easy
effective ways to detect and recognize different diseases and to use, don't need fully experienced men to handle, much
make it easy to examine and diagnose the actual problem. more accurate then the manually counting method and is very
Now a day's researchers are working on different computer economical way of cell counting for the countries like
vision applications for health industry. Such as segmentation Pakistan.
of kidney from ultrasound images, Cancer Detection Using
Pattern Recognition Technique, segmentation of brain images, 1.2 Importance of the Project
etc. The Dengue and malaria fever is very common now days in
Fast and cost-effective blood cell counting has major south Asia and specially in Pakistan. In Pakistan the dengue
importance in the medical world. The old conventional virus infected over 30,000 peoples (only in Punjab) in summer
methods of blood cell counting under the microscope render 2011. Blood related viruses are very common in south Asia
unreliable and unacceptable results and put an unendurable the most highlighted among them are Dengue and malaria,
amount of strain on the Clinical laboratory technicians. Even dengue fever is also known as break bone fever, when a
so there are latest hardware solutions such as the dengue virus infected mosquito bites a human the virus along
Heska'sHema'true hematology analyzer, but they are widely the mosquito saliva enter the skin of the person. This deadly
unavailable and very expensive machines and so the under virus then enters into the white blood cell of the patient and
developed countries like Pakistan are not resourceful to infected the cells, it damages the cell and its functionality and
provide such an expensive solution of blood cell counting in spread in the whole. Initially this virus can be detected in the
every hospital laboratory in the country. As a solution laboratory by investigating low white blood cell count, which
regarding this problem, this research based paper proposed a then leads to very low platelets (PLT) count ratio [1], and so
fast and cost effective software-based alternative method to blood test was the most common test to be diagnosed in the
count accurate blood cells. clinical laboratories

1. INTRODUCTION 2. LITERATE REVIEW

Here in this research based project we proposed a software An analysis of blood was exercised from far back to ancient
base solution related health industry which will assist the times. All three blood cell types performs its own role in
medical laboratory technician (MLT) to detect and find a healthy men's life and so count of different cell type of blood
blood cell count and produce an accurate cell count report. can identify different diseases that's the reason that complete
This would be very helpful to a physician in identifying the blood cell count is the most common test carried out in all
cause of his patient’s diseases. To count the blood cells in a clinical laboratories. Different techniques were practiced since
clinical laboratory different two methods and techniques are the discovery of blood cells in 1658. Before going into details
used. One is the old conventional method of cell counting of modern blood cell counting methods we should know the
under the microscope and the other is to produce cell counting history of cell counting and the developments in the
report by latest but very expensive haematology analyzer. But technology of cell counting which was finally implemented to
both these methods haveits own different drawbacks and quantification of the ingredients of blood
limitations.
2.1 Hematology
1.1 Limitations withExisting Methods In hematology we deal with the essentials of blood and the
The main problem with the method of counting manually tissues for the forming blood.[2] Hematology is used to
under the microscope is accuracy, this method needs a real identify and examine the cure for anemia, leukemia's and
experienced laboratory technician who is trained enough to hemophilia (a kind of blood disease). Hematological tests are
produce an accurate cell counting report, and even if the performed to check the results of certain treatments e.g.
laboratory technician if well trained and experienced still we cancer chemotherapy and also to get outcome about the
patients overall health.

18
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

2.2 History of Cell counting 2.4 Blood Cells sizes

Leeuwenhoek was the first person who attempted to count As mention in the table below that platelets are much smaller
blood cells using a glass capillary tube with graduation marks than RBCs and WBCs in size. A normal human body has a
of measured dimension and microscope to count. He selected platelets count of 150,000. Decrement in platelets count can
chicken to count red blood cells [3]. Afterwards, different result in body disorders and can cause disability of blood as a
techniques were introduced for diluting the blood which result can lead to death.
resulted in more accurate and easier counting using a shallow
rectangular chamber which had a thin cover glass and diluted
blood was injected into this glass. In the early 20th century a
Type of blood Size[u m]
technique using photoelectric device to count cells was
invented by Moldovan [4]. However, this attempt for cell white blood ce11 10-20
counting did not develop at that time because of the
unreliability of the photoelectric device. An automated blood- Red blood cell 6-10
cell counter technique was invented by Waiter H. Coulter [5]
in the mid 1950's for blood cell counting. The research was platelets 2-4
based on the technique known as “Coulter’s Principle” or the
Aperture Impedance technique. This technique uses the
resistivity of the blood cells because thee impedance of the 2-1: Size of blood cell types[6]
cells suspended in the diluting fluid is much more higher than
that of fluid was based on the fact that the resistivity of blood 2.5 Challenges for Modern cell Counters
cells is much higher than that of the diluting fluid. Most Counting of cells one at a time is the objective of cell counter.
modern cell counters serves on the basis of this extensively It may appear easy but it is more technically challenging then
developed since 1950’s. it seems. There arefew reasons which act as a challenge for
modern cells counter. The nature of the blood cells come first,
2.3 Cellular Elements of Blood secondly the difficulty in accounting for the volume of liquid
due to dilutions and finally the need to obtain a cell count
Different blood cell types are the main elements of the blood;
separately for each cell type. In the following few sections we
they are white blood cells (WBC), Red blood cells (RBC), and
will briefly describe the impact of these general problems.
the platelets (PLT).

2.3.1 Red Blood cell: Erythmytes 2.5.1 The Nature of Blood Cells
Small size and very concentration of the blood cells are the
Erythrocytes are the most important and major elements of
main technical challenges for it. For the separation of signals
blood. There are normally 4-6 million in number in a normal
from the cells, the counter must be able and it is also able for
human body. Hemoglobin a major part of RBCs, carry oxygen
checking of arising false signal from the cells and the removal
from the lungs to the tissues and carbon dioxide from the
of background noise. The sensing area of the device must be
tissues back to the lungs. If any variation in RBCs count is
small so that the particles cross it one at a time and large
found, it can result in many symptoms and diseases can attack
enough to allow the cell particles. Since blood cells are so
on an individual. So RBCs play an important role in
many in number so it may happens that more than one cell can
identifying a variety of disease.
pass through the sensing area at a single time and this is called
coincidence count and the error given by this type of counting
2.3.2 White Blood Cells : Leukocytes is called coincidence error. Dilution of sample and proper
WBCs are the minor part of blood cells as their count is 9,000 control at the parameters in sensor will minimize the
– 30,000 / mm3 for a newly born and after few weeks it coincidence error and will be the cause for reducing of
decreases to 6,000 – 11,000 / mm3. An adult has only 4,000 – coincidence. When cells dilution is done then cells passed one
11, 000 / mm3 of leukocytes. WBCs consist of neutrophils, by one from sensor which is very sensitive to detect traversal
basophiles, eosinophiles, monocytes and lymphocytes. The of cells.
lymphocytes control the immune system of human body and
fight against the harmful germs in the body. Lymphocytes
produce antibodies. Lymphocytes increase their number when
2.5.2 Sample Dilution
After a cell counting is completed, the volume came out after
a viral infection takes place. Neutrophils play a defensive role
concentration is same as the whole blood volume. Therefore
in attacking germs and harmful bodies. They also increase
the volume of all diluted solution must be kept so that it can
when bacterial infection is found in the body. The WBCs have
be refer back to the initial volume. One problem came in
a variety of life spans, some live few days and the others last
counting the volume that blood is in liquid form and it
for several of months. Leukocytes live in tissues and other
contains many particles of different densities so similarity of
parts of body but just use blood as a mean of transportation.
solution must be given a special attention since the cells can
settle to a surface. One other problem is the flexibility of red
2.3.3 Platelets :Thembocytes blood cells so that it makes the similarity in cells under many
Platelets are fragments of cytoplasm that are fired out in the flow conditions. Goldsmith proved it that after the
blood from large cells in the bone narrow. So some physicians deformation of red blood cells, cells try to mass in the
don’t consider them complete blood cells. Platelets work direction of cylindrical tube centre and make similarity in
importantly in blood clotting known as haemostasis. Vessel solution. So by using pipette one of the laboratory techniques
walls are surrounded by platelets to stop bleeding when distort concentrated cells when it is a homogenous solution.
injured. They also help in infections from enzymatic
reactions. In the commercial cell counter, the loosing track of the
relationship among the cells and original volume of the
sample must not be done for the required dilution. The

19
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

problem will be overcome by carefully analyzing the sample 2.7 Components involved in counting
and fluid. Complete system according to proposed method of blood cell
counting has a microscopic camera over a microscope,
2.6 Haemocytometer: Haemocytometer and a PC as shown below in the fig.
Haemocytometeris also known as new-bar counting chamber.
It can also be used in counting other type of microscopic
elements e.g. sperms, bacteria, fungus etc. Think glass
microscopic slide with a slip over to dispense the solution
equally in the chamber is specially design for counter
chamber. The chamber is fixed with a laser etched grid of
perpendicular lines, and blood cells are counter per unit
volume and the blood is in micro liter. Chamber is sensibly
crafted so area which is under the microscopic grids is easily
known and the depth of the chamber is also recognize by the
user so that he can easily find the solution’s volume for cells
counting.

Figure 2-5: Complete system for cell counting

3. METHODOLOGY
An adaptive automatic thresh holding technique based on
Otsu method has been proposed as a counting algorithm and
to enhance the accuracy of the microcell counting in
microscopic images the proposed algorithm is the famous
water shade algorithm. Following states the proposed
Figure 2-1: Haemocytometer algorithm for counting. An image captured by CCD camera
setup at microscope is fully transformed and then divided into
2.6.1 Representation of Haemocytomoter. cropped images of mxn blocks with known size. This image is
The formal internal representation of the grid’s numbering then converteded into binary image the images which are
represents in figure 2-4. The counting grid is composed of 9 microbial and below pre-specified pixels are considered noise
big squares, measuring 1 x 1 mm. In these squares, the in the binary image and are removed in the binary image.
centered square comprises twenty-five moderate size squares Morphological filters and the area opening are used for
for each one appraising 0.2 x 0.2 mm. This is additionally smoothing procedure.
separated into sixteen minute squares for each one measuring
out 0.05 x 0.05 mm. The big centurial square is known as the 3.1 Manual methods for cell counting
“erythrocyte” grid. The squares marked in red equate into
eighty microscopic squares, and are applied to count the PLT In Clinical Laboratory technicians prepare the slide by mixing
and RBC. The large squares highlighted in blue are related reign with known quantity of blood to examine it
accustomed to count the white blood cells in it under the microscope for blood cell counting. Count is
obtained by putting contented cells as inputs in different
equations.

3.1.1 Manual White Blood Cell (WBC) Counting

To manually count the WBC, 50 μl amount of blood is mixed
with dilution solution of amount 950 μl. The dilution result is
1:20. This counting causes lysis of RBC (i.e. cells are
destroyed by bursting), and staines the nucleus of the WBC.
After mixing the counting chamber is filled immediately. The
counting of the WBCs is started by the MLT after 2 mins in
the 4 large squares.
Using these parameters a formula has been derived for the
calculation of the WBCs:
i. In the big squares how much WBCs have been counted
ii. The cell solution has been diluted up to how much extent
iii. Total number of squares that are counted.
Figure 2-2: Schematic representation of the
Haemocytometer iv. Volume of one big square

(𝑖)# 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑊𝐵𝐶 𝑋 (𝑖𝑖) 𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛

(𝑖𝑖𝑖)# 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑠𝑞𝑢𝑎𝑟𝑒 𝑋 (𝑖𝑣) 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑜𝑛𝑒 𝑏𝑖𝑔 𝑠𝑞𝑢𝑎𝑟𝑒

20
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

3.1.2 Manually counted platelet (PLT) counting 3.3 Hematology AnalyzersLimitations

This automatic cell counting machine is widely unavailable.
The manual counting of PLT is very similar to that of the
The cost of latest automatic Hematology analyzer is very high
RBC counting. However contrary to the RBC counting the
for clinical laboratories in the countries like Pakistan.
RBCs arecompletely destroyed before analysis, the mixture of
dilution is 1:[Link] chamber is filled with suspension
after thorough mixing. To settle the platelets without the
chamber getting dried it is left for about 20-30 minutes in the
room temperature. The counting is just like red blood cells,
total 80 squares are counted that are small in size.
Using these parameters a formula has been derived for the
calculation of the PLTs:
i. . in the small squares how much PLTs have been counted
ii. The cell solution has been diluted up to how much extent
iii. total number of squares that are counted.
iv. volume of one small square

Figure 3-1Hematologyanlyzers
(𝑖)# 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑊𝐵𝐶 𝑋 (𝑖𝑖) 𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛
𝑖𝑖𝑖 # 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑠𝑞𝑢𝑎𝑟𝑒 𝑋 (𝑖𝑣) 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑜𝑛𝑒 𝑏𝑖𝑔 𝑠𝑞𝑢𝑎𝑟𝑒 3.4 Cell counting by proposed method
Proposed method in this paper a software that can be used for
counting all three types of blood cells accurately to diagnose
3.1.3 Manually counted Red (RBC) count the patients actual problem. Proposed method can perform all
To count RBC manaully, 10 μl amount of blood is diluted in standard cells counting which includes RBC, WBC and PLT
dilution solution of amount 1990 μl. The result of dilution is counts. The functioning in the proposed method is totally
1:200. The counting chamber is then immediately filled with based on computer vision technologies via image processing
suspension after the suspension has been thoroughly mixed. techniques. The input the software is an image of carefully
The RBCs take about approximately 3 minutes to get prepared slide of blood solution as described in above
settled,after that RBC counting is started by the MLT in 80 chapters, the input image is taken from a special type of
small squares. microscope camera attached to a very common compound
microscope found in every clinical laboratory. The proposed
Using these parameters a formula has been derived for the method is not about the preparation of slides having blood
calculation of the RBCs: solution that are to be viewed under the microscope. It still
needs a laboratory technician who takes blood from the
i. . in the small squares how much PLTs have been counted patient and prepares a slide as described in the above chapter.
ii. The cell solution has been diluted up to how much extent The input images that are taken through a microscope camera
are similar to what the laboratory technician observes under
iii. total number of squares that are counted. the microscope. Like hematology analyzer machine proposed
iv. volume of one small square method also need a laboratory technicians to take blood but
additional work in the method is to prepare the slide like they
do for manual counting method. The main purpose of the
proposed way of counting is to provide a cost effective
(𝑖)# 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑊𝐵𝐶 𝑋 (𝑖𝑖) 𝑑𝑖𝑙𝑢𝑡𝑖𝑜𝑛 method which is accurate and trustable as well as suitable for
underdeveloped countries like Pakistan. The proposed method
𝑖𝑖𝑖 # 𝑜𝑓 𝑐𝑜𝑢𝑛𝑡𝑒𝑑 𝑠𝑞𝑢𝑎𝑟𝑒 𝑋 (𝑖𝑣) 𝑣𝑜𝑙𝑢𝑚𝑒 𝑜𝑓 𝑜𝑛𝑒 𝑏𝑖𝑔 𝑠𝑞𝑢𝑎𝑟𝑒
is designed for the common men and is easy to use for the
end-user.
3.2 Draw Backs of manual count method
3.4.1 Functionality of proposed method
Manual inspection of images taken from microscope The overall Methodology of image processing techniques in
consumes a lot of time and is tiring. If the process of counting the proposed method is described in the figure bellow.
is disturbed, the MLT has to start the counting from beginning
An experienced MLT carries out the cell analysis by the
comparison of images of cell types she sees and is familiar
with. On the other hand less experienced MLT in order to
confirm on the cell types would have to check with medical
manuals repeatedly to make sure that the counting of given
sample is accurate. Thus manual counting methods are
vulnerable to human mistakes that can easily result in errors. Figure 3-2:over all methodology of image processing
After analysis of slides of blood cell, they are taken away and First of all an input image of already prepared slide is take
kept. The method of retrieving and analyzing the sample for from the camera attached to the microscope, then through
future usage is not possible with manual count. different image processing techniques and image filters the
extra unwanted information is removed from that input image.
After removal of noise the area of interest that is the area

21
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

where we need to count the blood cells are cropped from the
image and by labelling algorithm of connected pixels blood
cells can be count to put in different equations to abstained
cell count report.

3.4.2 Image Acquisition

A total of 45 images of all three types of cell in it were
collected from Kohat University of science and technology
(KUST) zoology and microbiology department. The images
of blood are taken from microscope camera as described
above. The images taken were verified by the microbiologist
and homology experts to make sure that we have taken correct
snap shots input to put methods. The images includes three
imagesof each blood test three for each and 45 images in
total.

3.4.3 Developing the Image Processing

algorithms
Matlab R2010a is used in this method of blood cells counting.
The main reason of using Matlab R2010 for the method is to
make a software based solution which is cost effective and
suitable for countries like Pakistan to be implemented on
large-scale in the medical world. The development of
proposed method has different phases. This enrolled the pre-
processing techniques of image which has to be followed
before the blood cell counting in the input [Link] of
pre-processing applied techniques on the image is contrast
enhancement for better result to convert it to greyscale
image, and then finally to binary image for easy and accurate
counting for the filtered blood cells in the input image for this
method the technique applied is shown in the figure 9 bellow.
Considering the bellow flow chart the input image will be Figure 3-3: Flow chart of image processing
converted to the gray scale in very first step. Then relying on
the grayscale values of the image, the new threshold values 3.4.6 Histogram and Thresholding
through the well known otsu method will be determined for
different blood cell counts that is RBC and WBC. These When the original coloured (RGB) image of blood is
values obtained from the auto threshold method will be used converted into grayscale intensity values of pixels, the gray
to remove noise from the input image and filter up the level histogram can represent the distribution of pixel
remaining noiseless blood cells .then the resulted image after intensities of that image. We can easily see the frequencies of
these filtering processes it is then converted to the binary all intensity, and on bases of gray level values in the
image to calculate the blood cells. But If still these images has histogram the image can be analysed. The histogram study of
noise ratio in it then above thresholding and segmentation gray level intensities is important to make a line between
techniques will be applied again and again until we get and blood cells and the background. Between the two peaks of the
noise free clear binary image having cells in it. Flow diagram histogram the threshold value lies some where. In Matlab
of the image is as under. R2010a software, IMHIST function can be used to develop
the histogram of an image. Fig. 10 is a histogram example of
3.4.4 Pre-processing of input image blood cell image. The obtained histogram is divided into
Grayscale image, contrast enhancement in the image and to three different sections one section for RBC, the other is for
convert that refined image into binary image are the pre- WBC in the blood and third the background of these blood
processing image techniques used in the making of this image cells indicated by (a), (b) and (c) separately. And so the
processing system. threshold value lies between the area (a) and (b) that is
between the peaks of WBC and the RBC. To segment the
3.4.5 Gray-scale Image region of interest this threshold value is used. Basically the
method of converting gray scale image into the easily handled
Initial input image that is taken from the microscope camera is binary image is called thresholding of the image. The
a color image. to make the image processing easy for the advantage of thisthresholding is to separate the blood cells
proposed method development , the original coloredinput from the back ground. The thresh Threshold value obtained
images will be converted into grayscale images from the from the histogram is applied on every filtered image to get
range starting form 0 (black pixel) to 255 (pure white pixel) . the refined blood cells for counting., we placed the threshold
Pixel intensity in the gray scale image can be represented In values into following limits;
Matlab R2010a, and this conversion can be done by using a
built-in image processing tool of matlabR2010 RGB2GRAY a) Pixel intensity value in the blackish area of 0 to 89
function. The RGB2GRAY converts the coloured RGB image are converted to pure white 255 pixel value this
into grayscale pixel intensities starting from 0-255. area represents the white blood cells,

22
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

b) Pixel intensity value in the gray area of 90 to 170 different views. Also pathologist can get the result of blood
are converted to pure white 255 pixel value this cell test within 20 seconds to 1 min. as tested.
area represents the Red blood cells
c) Pixel intensity value in the blackish area of 171 to 4.1 Obtained Results
255 are converted to pure black 0 pixel value this We have collected more than 15 blood samples for
area represents the background experiments and finally tested these samples by all methods of
cell counting including our proposed method and we have
found that our results are 95 plus % accurate.

4.1.1 Sample number 1

Results comparison table of the 1st tested sample for White
blood cells is as under.

Manual Hematology Proposed

method analyzer method
10.2 11.4 11.7
Table 4-1: Results of first sample
Figure 3-4: histogram of binary image
Results of WBC of first blood sample is
3.4.7 Conversion to Binary Image
Binary images are those images having only two pixel  Manual counting = 10.2K per micro liter
intensity values, zero (pure black) and 1 (white). By  Hematology analyzer = 11.4K per micro liter
thresholding the unwanted area in color images or images  counting by proposed method = 11.7K per micro liter
with gray level pixel intensity binary images can to formed. In
this proposed method development with the help of threshold
value obtained from the histogram peaks the binary images is
formed to separate white and red blood cells. if the result is
satisfied the pre-processing phase completes here. but if more
filtration id needed the image has to passes from all above
processes again until we obtained our required image.

3.4.8 Image Enhancement

In the first method, image enhancement technique will be
applied after the blood cell image has been converted into
binary image. Normally, the original binary image consists of
small spots within the image. These small spots are assumed
as noise and need to be removed from the image. Hence, we
proposed two image enhancement methods to remove the Table 4-2: Print of results by both methods
noise; i) removing pixels technique and ii) Gaussian filter. By From these results we have seen that the proposed method is
using the removing pixels technique, every object in an image working and results very close to the counting results by
which contains less than 100 pixels will be eliminated. This is hematology analyzer.
the best value that can be used because it does eliminate only
the unwanted noise in the image. In Matlab Ra2010a, this can
be done by using BWAREAOPEN function. Besides that,
Gaussian filter was also applied on the binary image of the
blood cell. Gaussian filter removes noise by smoothing but
also blurs the image. Gaussian provides gentler smoothing
and preserves edges better than a similarly sized mean filter.
In the second method of image processing, the image
enhancement technique was applied on the original image.
For this method, only Gaussian filter was implemented on the
original image.

4. RESULTS AND DISCUSSION

This system has implemented the segmentation part and after
testing this system it is concluded that, it is more timely
efficient than the existing systems. It is reliable and cost
effective than automated method. The aim of this system is to
provide CBC i.e. completeblood Cell Count, which has been
Figure 4-1: Graph of 1st sample
achieved by using the Powerful Image processing technique.
4.1.2 Sample number 2
This system is more efficient in reducing the valuable time results comparison table of the second tested sample for
than manual system. The system is user friendly so that the White blood cells is as under.
pathologist can observe the blood cell sample image in

23
 International Journal of Computer Applications (0975 – 8887)
Volume 50 – No.1, July 2012

Table 4-3: Results of second sample 5. CONCLUSION AND FUTURE

Results of WBC of first blood sample is RECOMMNDATION
In this paper we have developed a software based solution for
blood cell count for underdeveloped countries like Pakistan
Manual Hematology Proposed
which are resource less to produce and provide expensive
method analyzer method hematology analyzer in every hospital laboratory of the
9.8 8.3 8.35 country.
 Manual counting = 9.8K per micro liter
 Hematology analyzer = 8.3K per micro liter Proposed method of cell counting is fast, cost effective and
 counting by proposed method = 8.35K per micro liter accurate to produce blood cell report. it gives 95-100%
accurate results for white blood cell count which an important
test carried out to diagnose diseases like dengue at initial
stages. but due to limited resources we were unable to find
some good images of platelets and red blood cells but the idea
and procedure for these tests are also same like it is for white
blood cell counting.

5.1 Future recommendation

Proposed method for cell counting white blood cells in the
paper is similar for counting remaining two types of cell
counting. and through the same procedure we can develop an
other algorithm for detecting malaria parasite in the blood,
malaria parasite also found in blood and the current method of
detecting malaria parasite is to prepare the slide and study is
under microscope to recognize malaria parasite in the blood,
Figure 4-2: Graph of 2nd sample by improving our proposed algorithm we can also detect
malaria parasite through computer vision technologies and
from these results we have seen that the proposed method is techniques of image processing techniques. but this procedure
working and results very close to the counting results by would need a little advance microscope then the ordinary used
hematology analyzer. in the clinical labs to achieve 100% accurate results.

6. REFERENCES
[1] Graham Ramsay, Commercial biosensors, John Wiley &
Sons, Inc, New York, 1998. 1-Stat Corporation official
web page, [Link] [Link]/ , 1999.
[2] Graham Ramsay, Commercial biosensors, John Wiley &
Sons, Inc, New York, 1998. 1-Stat Corporation official
web page, [Link] [Link]/ , 1999.
[3] Hajdu, SI. The discovery of Trichomonasvaginalis. Act
Cytologica 1998;42:1075.
Table 4-4: Print of results by both methods
[4] Bennett, JH. The Employment of the Microscope
4.2 overall results inMedical Studies. Stewart, Edinburgh, 1841.
[5] Hajdu, SI. The discovery of Trichomonasvaginalis.

Overall counting [6] JH. The Employment of the Microscope in Medical

Studies. Stewart, Edinburgh, 1841.
comparison [7] Bennett, JH. Case of hypertrophy of the spleen and liver,
in which death took place from suppuration of the blood.
15
10 Manual
method
5
0 Hematology
analyzer
1 2 3 4

Table 4-12: Graph show comparison of overall cell

counting methods
above graph shows that manual way of blood cell counting
has greater chance of error in it while our proposed method of
cell counting and counting on hematology analyzer are totally
or almost similar.

24

View publication stats