Bio medical image segmentation using marker controlled watershed algorithm a case study

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 03 Special Issue: 03 | May-2014 | NCRIET-2014, Available @ http://www.ijret.org 26
BIO-MEDICAL IMAGE SEGMENTATION USING MARKER
CONTROLLED WATERSHED ALGORITHM: A CASE STUDY
Ravi S1
, A M Khan2
1
Research Scholar, Department of Electronics, Mangalore University, Mangalagangothri – 574 199
2
Professor and Chairman, Department of Electronics, Mangalore University, Mangalagangothri – 574 199
Abstract
Image segmentation is a process in which the objects of interests are selected and separated from the total image to analyze and
extract the required data. It is one of the image processing methods, which is widely open for research. Many algorithms have been
proposed till date to improve the quality of segmentation and the best one is yet to design. Watershed algorithm is one such reliable
approach to segment homogeneous regions in an image. It is superior to region growing method in terms of number of regions
obtained after processing. In this paper a robust approach is discussed to improve, the primary drawback of watershed algorithm,
over-segmentation. We discuss morphological marker controlled method and the use of gradient image for watershed algorithm to
minimize over-segmentation.
Keywords: Watershed Algorithm, Morphological Operation, Marker Controlled Watershed, Morphological
Reconstruction, Image Segmentation.
------------------------------------------------------------------------***---------------------------------------------------------------------
1. INTRODUCTION
Image is two-dimensional light intensity function of spatial
coordinates. If the intensity function and the spatial
coordinates are finite and discretised, the image is said to be
digital in nature. Processing of such digital image using a
digital computer is called digital image processing[9]. There
are three levels of image processing like low-level, mid-level
and high level image processing. All these levels are
concerned mainly in improving the quality of image for
human visualization and to extract more information from the
image. Among these, image segmentation is one such method
which has been widely explored to enhance the quality of
information available in the image[6].
Image segmentation can be done in three distinct ways; based
on the intensity values (threshold based method), based on
discontinuities in the image (edge based method), based on
homogeneous regions present in the image (region based
method)[10]. The said methods are the conventional methods
used for segmentation. Among these, edge based method and
region based methods are widely tested. Edge based method
identifies the edges, which are basically sharp intensity
transitions, local to a region, in the images and mark the pixel
as boundary element. In contrast, region based method groups
the pixels which share common characteristics to form
homogeneous regions[1]. Region based segmentation can be
done in two different ways as growing method and split &
merge method. Region growing method is usually preferred
for its reliable and robust output.
Watershed algorithm is hybrid approach of „good
characteristics‟ of edge based segmentation and region based
segmentation[9] and hence gives better result than its parents.
In this paper we discuss the theory of watershed transform and
apply the algorithm for few test images for better
understanding. We also discuss marker controlled approach
using morphological operations to improve the quality of
watershed algorithm and to overcome its drawback of over-
segmentation.
Morphological operations are the mathematical set theory
operations that are very simple to use and mainly concerned
with the ordering of the elements instead of the numerical
values of members of the set[6]. This mathematical
morphology when combined with the logical operations gives
very good results in image processing for performing
operations like blurring, sharpening reconstruction and so
on[8]. Morphological operations are very much useful in
watershed algorithm to construct dam between neighboring
regions which is the key process in watershed algorithm.
The paper is arranged as follows: section II highlights
watershed algorithm, section III is on basic morphological
operations, in section IV we discuss marker controlled
watershed algorithm, in section V application of our algorithm
is shown and in section VI results are discussed.
2. WATERSHED ALGORITHM
The watershed algorithm includes the three basic segmentation
approaches like threshold based, edge detection and region

__________________________________________________________________________________________
based segmentation[1][2]and hence it provides more stable
results than these methods individually. It uses morphological
operations as well to set proper boundaries between adjacent
regions and morphological operations are discussed in next
section.
To understand watershed algorithm it is to be known the
following three „points‟:[9]
1. Points belonging to regional minimum
2. Point of catchment basin / watershed lines.
3. Points at which water drop can more likely to occupy
other than a specific watershed.
The watershed lines are those single minimum values to which
a drop of water falls with certainty. The gradient of the image
is to be found before applying watershed. The characteristic of
a pixel will be compared with the neighboring pixel and if
found similar, the pixels are added to form a region. The
process is carried out till edge of the region is found or the
neighboring regions are above to merge. At this stage a dam is
built to avoid the merging of two different regions. As a result,
after testing all the pixels only the top of the dam is visible and
is the resultant of segmentation of various regions in the test
image.
Let M1, M2, …,MR sets denoting the coordinates in the
regional minima of an image g(x, y), where g(x, y) is the pixel
value of coordinate (x, y). Let C(Mi) be the coordinates in the
catchment basin associated with regional minimum Mi and let
T[n] be the set of coordinates (s, t) for which g(s, t) < n and is
given by
T[n]= {(s ,t) | g(s, t) < n}
Then the process of watershed algorithm is discussed as
below:
Step1:The boundary value of the pixels is to be found and
the minimum value is to be noted
The boundary values of the pixels of g(x,y) is to be found and
the minimum value is to be assigned to Mi. Flooding is done
by initializing n=min+1
Let Cn(Mi) as the coordinates in the catchment basin
associated with minimum Mi that are flooded at stage n.
Step2:Compute catchment basins.
Compute Cn(Mi) = C(Mi) ∩T[n]
Cn(Mi)
=
1 𝑓𝑜𝑟 𝑥, 𝑦 ∈ 𝐶 𝑀𝑖 𝑎𝑛𝑑𝑎𝑙𝑠𝑜 𝑥, 𝑦 𝑏𝑒𝑙𝑜𝑛𝑔𝑠𝑡𝑜𝑇[𝑛]
0 𝑜𝑡𝑕𝑒𝑟𝑤𝑖𝑠𝑒
Let C[n] denote the union of the flooded catchment basins at
stage n:
C[n] = Cn(Mi)
𝑅
𝑖=1
Set n = n + 1.
Step3:Derive the set of connected components
Derive the set of connected components in T[n] denoting as Q.
For each connected component q ∈ Q[n], there are three
conditions:
a. If connected components is empty it represents a
new minimum is encountered.
If 𝑞 ∩ 𝐶 𝑛 − 1 is empty, connected component
q is incorporated into C[n - 1] to form C[n]
because it represents a new minimum is
encountered.
b. If connected components contains at least one
connected component it means connected
components lies within the catchment basin of
some regional minimum.
If 𝑞 ∩ 𝐶 𝑛 − 1 contains one connected
component of C[n - 1], connected component q
is incorporated into C[n - 1] to form C[n]
because it means q lies within the catchment
basin of some regional minimum.
c. If connected components contains more than one
connected component it represents all or part of
a ridge separating two or more catchment basins
and set them as “dam”.
If 𝑞 ∩ 𝐶 𝑛 − 1 contains more than one
connected component of C[n- 1], it represents all
or part of a ridge separating two or more
catchment basins is encountered so that we have
to find the points of ridge(s) and set them as
“dam”.
Step4. Construct dam for all possible catchment basins.
Construct C[n] using the values obtained for Cn(Mi) and c[n]
Set n = n + 1.
Step5. Repeat Step 3 and 4 until n reaches max + 1.
3. MORPHOLOGICAL OPERATIONS
Mathematical morphology is a tool for extracting region
shape[3]. It can also be used for pre-processing operations
much like filters used in spatial filtering operations.
Morphological operations are centered on set theory but, when
used along with logical operations it makes it much simpler

__________________________________________________________________________________________
than the conventional methods like convolution used in image
processing[6]. It is based on ordering of the pixels instead of
the numerical value of pixels. Morphological operations uses
binary image of very small size called as structuring elements
to process the test image. The structuring element works much
like the spatial filters that are used in spatial domain
processing[8]. Dilation and Erosion are the two fundamental
operations that are used in morphological operations and any
other morphological operation, like opening, closing,
reconstruction etc., are the combination of these two
operations[9].
Dilation: The dilation of an image A by structuring element B
is defined as
𝐴 ⊕ 𝐵 = 𝑧| (𝐵) 𝑧 ∩ 𝐴 ≠ ∅
If set B is reflected about its origin and shifted by z, then the
dilation of A by B is the set of all displacements z such that B^
and A have at least one common element. The dilation
operation makes the boundary of the object to grow, to the
extent decided by the shape and size of the structuring
element. This effect is analogous to the smoothing operation
performed in the spatial low pass filtering application. This
operation is used to fill „holes‟ in the object of an image. The
other consequence of performing the dilation operation is the
blurring effect.
Erosion: The erosion of an image A by structuring element B
is defined as
𝐴 ⊖ 𝐵 = 𝑧| (𝐵) 𝑧 ⊆ 𝐴
The erosion of image A by structuring element B is the set of
all points z such that the structuring element B is translated by
z is a subset of the image. The erosion operation is
contradictory to the dilation operation and it causes the
boundary of the object to shrink and is decided by the shape
and size of the structuring element. The thinning effect is
equivalent to the spatial high pass filtering. The erosion
operation removes those structures which are lesser in size
than that of the structuring element. It can be used to remove
the noisy „connection‟ between two objects in the image.
Image reconstruction: Reconstruction is a morphological
transformation that involves two images, a marker image and
a mask image, and a structuring element.
If G is the mask and F is the marker, the reconstruction of G
from F, denoted RG(F), is defined by the following iterative
procedures:[8]
1. Initialize h1to be the marker image, F.
2. Create the structuring element: B
3. Repeat:
h k+1 = (h k⊕ B)∩ G
until h k+1 h k .
4. R G (F) = h k+1.
Marker F must be a subset of G: F⊆G
4. MARKER CONTROL METHOD:
The application of watershed algorithm is greatly influenced
by the noise and other local irregularities in the image [5].
This leads to over-segmentation of regions to an extent that
the segmented image itself looks like a noisy image most of
the times and hence can be useless for diagnosing the
abnormality by a doctor. This is the major drawback of
watershed algorithm [6] [5][3][9]. To avoid this problem
markers are used for segmentation. A marker is a connected
component belonging to an image. The markers which are
connected components do possess the same intensity values
and are treated as regional minima. Markers can be classified
as internal (foreground) or external (background) depending
on its location from region of interest.
The use of markers gives a priori knowledge about
segmentation; this knowledge often consists in the number and
the positions of the regions through the definition of some
markers, and thus simplifies the problem of over-
segmentation. The result of using markers is shown in section
VI which clearly shows the improvement of segmentation
over conventional watershed method.
5. IMPLEMENTATION
We designed the algorithm by combining the watershed
algorithm, morphological operations and by providing markers
for better segmentation. The difference between the existing
approach and ours is that we had used bigger size gradient
filters compared to the filters used for conventional watershed
algorithm. The methodology followed is given below.
Fig-1: Flow chart of the proposed algorithm.
6. RESULTS AND DISCUSSION
We had applied the above algorithm on MRI of brain and on
mammogram images and compared with the conventional
watershed algorithm and the result is given below. The first
Read the image and convert it into suitable form
Find out the gradient of the image using a bigger
filter
Apply morphological image reconstruction
Calculate regional maxima in image
Obtain foreground markers & background
markers
Apply watershed algorithm to get segmented
result

__________________________________________________________________________________________
column of image is the original test images; of which two
corresponds to MRI of brain and other two are mammogram
images. Second column of images is the output of
conventional watershed algorithm, third column corresponds
to output of the proposed method and fourth column is the
result of overlapping third column on the first column. The
results shown below are obtained using Matlab2013a with
system configuration Inter(R) Pentium(R)4 ,CPU 2.80GHz,
1GB RAM, 32 bit OS.
It is evident that the proposed algorithm gives better result
than the conventional algorithm. The problem of over
segmentation is overcome by applying markers on the gradient
image. By declaring the internal and external markers we
restrict the segmentation to a single object and the
background. To carry out this segmentation we can use any of
the approaches available but the watershed gives better result.
Fig-2 First column: Original images; Second column: Conventional watershed output; Third column: Result of proposed algorithm;
Fourth column: Output of proposed algorithm overlapped with original image.
original image Waterhed Applied before smoothing

__________________________________________________________________________________________
We conclude that the proposed algorithm is superior to
conventional algorithm in terms of over-segmentation. Despite
of giving a good output the proposed algorithm in future it can
still be improved to get accurate results by using mathematical
transformations.
7. FUTURE WORKS
We have observed that though there is considerable reduction
in the over-segmentation, the obtained image is not ready-to-
use for diagnosing since it fails to mark the boundary between
adjacent regions that vary slightly in the intensities. One of the
possible ways to overcome this is by extracting high frequency
components in the images to give precise idea about markers.
We are working on this to make the processed image ready to
use at once the output is obtained.
REFERENCES
[1] Kostas Haris, Serafim N. Efstratiadis,
NicosMaglaveras, and Aggelos K. Katsaggelos,
“Hybrid Image Segmentation Using Watersheds and
Fast Region Merging”, IEEE Transactions On Image
Processing, Vol. 7, No. 12, December 1998.
[2] S. Beucher,,”The Watershed Transformation Applied
to Image Segmentation”, Centre de
MorphologieMathematique, Ecolos des Mines de Paris,
France.
[3] K. Parvati, B. S. Prakasa Rao and M. Mariya Das
Image Segmentation Using Gray-Scale Morphology
and Marker-Controlled Watershed Transformation
Volume 2008, Article ID 384346, 8 pages, Hindawi
Publishing Corporation Discrete Dynamics in Nature
and Society.
[4] SébastienLefèvre ,“Knowledge from Markers in
Watershed Segmentation” LSIIT, CNRS / University
Louis Pasteur - Strasbourg I Parcd‟Innovation, Bd
Brant, BP 10413 67412 IllkirchCedex, France.
[5] Amanpreetkaur, Ashish Verma, “The Marker-Based
Watershed Segmentation –A Review” International
Journal of Engineering and Innovative Technology
(IJEIT), Volume 3, Issue 3, September 2013.
[6] Ravi S and A M Khan, “Morphological Operations for
Image Processing: Understanding and its
Applications”, Proc. 2nd National Conference on
VLSI, Signal processing & Communications
NCVSComs-2013, 11 -12th Dec. 2013.
[7] Sameer Ruparelia, Implementation of Watershed Based
Image Segmentation Algorithm in FPGA, Master
Thesis Nr. 3256, Department of Parallel Systems,
Institute of Parallel and Distributed Systems,
Universit• atsstra, Stuttgart.
[8] Marker Controlled Watershed Algorithm IPT, MatLab
Documentation.
[9] R.C Gonzalez and R.E. Woods, Chapter 9& 10;
„Digital Image Processing‟, third edition, PHI
Publications, 2008.
[10] S. Nagabhushana, “Computer Vision and Image
Processing”, New Age International Publishers, 2005.
[11] http://www.mathworks.in

Bio medical image segmentation using marker controlled watershed algorithm a case study

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Bio medical image segmentation using marker controlled watershed algorithm a case study (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Bio medical image segmentation using marker controlled watershed algorithm a case study