Вы находитесь на странице: 1из 7

International Journal of Innovations & Advancement in Computer Science

IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

Detection and Percentage Calculation of Leukemic Cells in Blood


Smear using Image Processing
Tarak Das1, Saumadritaa Kar2, Arnabesh Das2 & Farhana Afreen2
1
Assistant Professor,2 B-Tech Final Year Students
Department of Biomedical Engineering,
Netaji Subhash Engineering College,Garia, Kolkata

ABSTRACT
Leukemia, also spelled leukaemia, is a group of cancers that usually begin in the bone marrow and result in increased
number of abnormal white blood cells. These white blood cells are not fully developed and are called blasts or leukemic
cells. Symptoms may include bleeding and bruising problems, feeling tired, fever, and an increased risk of infections.
Blood tests or bone marrow biopsies are typically used for diagnosis. However manually checking for abnormal cell
shapes in microscopic images is extremely time consuming and lacks accuracy. Our aim is to thus detect the leukemic
cells from blood smear images and calculate the percentage of abnormal cells using MATLAB based image processing
techniques. This work is still in its infancy. Progress of the work will identify the stage of cancer for leukemic patients by
simply processing a blood smear image of the patient.

Keywords: Leukemia; Blood Smear; MATLAB; Image Processing; Algorithm; Detection; Leukemic Patient

1. INTRODUCTION
Clinically and pathologically, leukemia is subdivided into a variety of large groups.The first division is
between its acute and chronic forms:
Acute leukaemia [1] is characterized by a rapid increase in the number of immature blood cells. The crowding
that results from such cells makes the bone marrow unable to produce healthy blood cells. Immediate
treatment is required in acute leukaemia because of the rapid progression and accumulation of the malignant
cells, which then spill over into the bloodstream and spread to other organs of the body. Acute forms of
leukaemia are the most common forms of leukaemia in children.
Chronic leukaemia [2] is characterized by the excessive build-up of relatively mature, but still abnormal, hite
blood cells. Typically taking months or years to progress, the cells are produced at a much higher rate than
normal, resulting in many abnormal white blood cells. Whereas acute leukaemia must be treated immediately,
chronic forms are sometimes monitored for some time before treatment to ensure maximum effectiveness of
therapy. Chronic leukaemia mostly occurs in older people [3], but can occur in any age group.
Primary objective of this project is to detect leukemic cells by Circular Hough Transform [4], [5] & calculate the
percentage value of the abnormal cells among non-infected cells.
Here we have taken 2 images, one with leukemic cells another a normal blood smear and used MATLAB as
the image processing software. The two images were acquired, converted to gray scale, circles were detected
using Hough transform and the percentage [6] was found out using the following formula:

Here, what we designed was a generalized algorithm to detect and

236 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

calculate percentage of abnormal cells, which will help us to carry the research forward and establish a range
of percentages to detect malignancy.

2. AIM AND OBJECTIVE OF THE WORK


The aim of the present work is to count leukemic cells with respect to normal cells, in percentage form, by
using a blood smear image. The sample images are processed in MATLAB, followed by the detection of
abnormal cells with a high radius value. By using the data available in MATLAB workspace, we detect the
abnormal cells using our own algorithm.

3. PROGRAM DEVELOPED FOR DETECTION OF ABNORMAL CELL


A = imread('Abnormal1.jpg');
subplot(3,3,1);
imshow(A);
title('Original');
I = rgb2gray(A);
subplot(3,3,2);
imshow(I);
title('Grayscale');
I = wiener2(I, [5 5]);
subplot(3,3,3);
imshow(I);
title('Weiner Filtered');

%Enter the radius range for all cells


x = input('What is the minimum radius?');
y = input('What is the maximum radius?');
[centers, radii] = imfindcircles(I,[x y],'Sensitivity',0.9,'Edge',0.15,'ObjectPolarity','dark');
subplot(3,3,4);
imshow(I);
hold on
viscircles(centers, radii,'EdgeColor','b','LineWidth',2);
title('Detection of All Cells');
All = numel(radii);

% Detect abnormally large cells by changing the edge threshold factor to a


% higher value
[centers,radii] = imfindcircles(I,[x y],'Sensitivity',0.9,'Edge',0.4,'ObjectPolarity','dark');
subplot(3,3,5);
imshow(I);
hold on
viscircles(centers, radii,'LineStyle','-','LineWidth',2);
title('Detection of Abnormal Cells');
abnormal = numel(radii);
Num = ((abnormal/All)*100);
fprintf('You have just found %f%% Abnormal Cells in the image.\n',Num);
%// Display message box
Msg = sprintf('You have just found %f%% Abnormal Cells in the image.\n',Num);
msgbox(Msg)
237 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

4. DEVELOPMENT OF AN ALGORITHM FOR DETECTION OF LEUKEMIC CELL IN


PERCENTAGE FORM
Sensitivity factor is the sensitivity for the Circular Hough Transform accumulator array, specified as the
comma-separated pair consisting of 'Sensitivity' and a nonnegative scalar value in the range [0, 1]. As we
increase sensitivity factor, imfindcircles will detect more circular objects, including weak or obscure circles
(increases false detection). Sensitivity is by default 0.85.
Edge sets the range of threshold for determining the edge pixels, specified as the comma-separated pair
consisting of 'EdgeThreshold' and a nonnegative scalar value in the range [0,1]. We specify 0 to set the
threshold to zero-gradient magnitude and 1 to set the threshold to the maximum gradient magnitude. The
function imfindcircles detects more circular objects (with both weak and strong edges) when you set the
threshold to a lower value. It detects fewer circles with weak edges as you increase the value of the threshold.
By default, imfindcircles chooses the edge gradient threshold automatically using the function graythresh.

FIG 1: ALGORITHM FOR DETECTING AND CALCULATING THE PERCENTAGE OF ABNORMAL


CELLS IN AN IMAGE OF blood SMEAR

238 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

Object Polarity indicates whether the detected objects are darker or lighter than the background. viscircles
(centres, radii) draws circles with specified centres and radii onto the current axes. The function n = numel
(A) returns the number of elements, n, in array A, equivalent to prod (size (A)). The algorithm is depicted in
Fig: 1. Fig: 2 represents the process followed to detect percentage of abnormal cells.

5. PROCESS CHART

Image Acquisition

RGB to Grey

Detection of All cells using Detection of Myoblast


Circular Hough Cells using Circular
Transform Hough Transform

Adjustment of Sensitivity Adjustment of Sensitivity


& Edge Threshold Values &Edge Threshold Values
for Proper Detection for Proper Detection

Percentage Calculation of Myeloid


Cells in the Given Image:

No. of Abnormal Cells

* 100%
Total No. of Cells

Display Message
FIG 2: PROCESS CHART TO DETECT THE PERCENTAGE OF ABNORMAL CELL

239 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

6. RESULTS AND DISCUSSION


First, we have taken a blood smear image of a leukemic patient and used our proposed program to detect the
number of abnormal cells in percentage form. The image is dragged into work space to import it, as depicted
in Fig: 3. Then we have entered the minimum and maximum value of R that serves as the limiting range of
radius for detecting abnormally sized cells. Here we have used Rmin=12 and Rmax =35 for the sample image
Abnormal1. Fig: 4, shows the output with step by step image formation.

FIG 3: OUTPUT OF THE RESULT TO CALCULATE PERCENTAGE OF ABNORMAL CELL

FIG 4: IMAGES TO DETECT ABNORMAL CELLS

240 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

The results of same process followed for a normal blood smear image is depicted in Fig:5 and Fig:6. where we
have used same input value Rmin=12 and Rmax =35 and shown that the percentage of abnormal cell is zero.

FIG 5: OUTPUT OF THE RESULT TO CALCULATE PERCENTAGE OF ABNORMAL CELL

FIG 6: IMAGES FOUND FOR NORMAL CELLS USING THE SAME PROGRAM

7. CONCLUSION:
Finally, we can conclude that the proposed algorithm or program, can not only perfectly identify the leukemic
cells from a microscopic blood smear image of a leukemia patient, but can also give perfect results for a

241 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen
International Journal of Innovations & Advancement in Computer Science
IJIACS
ISSN 2347 8616
Volume 6, Issue 5
May 2017

similar blood smear of a normal patient. This program can perfectly detect the leukemic cell in percentage
form with respect to a normal blood smear.

8. FUTURE SCOPE OF THE WORK


Future scope of the present work is enormous. We can acquire a number of blood smear images for both
normal and leukemic patients from a cancer research institute. The images must be taken from same optical
instrument. After image acquisition, the sample images can be processed using our algorithm. By performing
a statistical analysis of the output, a final threshold value can be found, that will be able to distinguish between
leukemic and normal images. It would thus become easier for us to detect if a patient is leukemic, by simply
taking the image of his or her blood smear and running it through the proposed algorithm.

9. REFERENCES
1. Shrutika Mahajan, Snehal S. Golait, Ashwini Meshram, Nilima Jichlkan, Review: detection of types of acute
leukemia, IJCSMC, Vol. 3, Issue. 3, March 2014, pg.104 111.
2. Ivan Pasic, Jeffrey H. Lipton, Current approach to the treatment of chronic myeloid leukemia, April 2017Volume 55,
Pages 6578.
3. John G Gribben, Chronic lymphocytic leukemia: planning for an aging population, Expert Rev Anticancer Ther.
2010 Sep; 10(9): 13891394., doi: 10.1586/era.10.127.
4. Subhan, Ms. Parminder Kaur, Significant Analysis of Leukemic Cells Extraction and Detection Using KNN and
Hough Transform Algorithm, International Journal of Computer Science Trends and Technology (IJCST) Volume
3 Issue 1, Jan-Feb 2015.
5. Mausumi Maitra , Rahul Kumar Gupta, Manali Mukherjee, Detection and Counting of Red Blood Cells in Blood
Cell Images using Hough Transform, International Journal of Computer Applications (0975 8887) Volume 53
No.16, September 2012.
6. Grzegorz S. Nowakowski, James D. Hoyer, Tait D. Shanafelt, Clive S. Zent, Timothy G. Call, Nancy D.
Bone, Betsy LaPlant, Gordon W. Dewald, Renee C. Tschumper, Diane F. Jelinek, Thomas E. Witzig, and Neil E.
Kay, Percentage of Smudge Cells on Routine Blood Smear Predicts Survival in Chronic Lymphocytic Leukemia, J
Clin Oncol. 2009 Apr 10; 27(11): 18441849.Published online 2009 Mar doi: 10.1200/JCO.2008.17.0795

242 Tarak Das, Saumadritaa Kar, Arnabesh Das & Farhana Afreen

Вам также может понравиться