IJARCCE
ISSN (Online) 2278-1021
ISSN (Print) 2319 5940
International Journal of Advanced Research in Computer and Communication Engineering
ISO 3297:2007 Certified
Vol. 5, Issue 12, December 2016
Implementation of Watershed Segmentation
Lalitha K1, Amrutha R1, Stafford Michahial1, Dr M Shivakumar2
Asst Professor, Dept of IT, GSSSIETW1
Prof & Head, Dept. of EIE, GSSSIETW 2
Abstract: This paper focuses on marker based watershed segmentation algorithms. As marker based watershed
segmentation algorithm causes over segmentation and cause noise in the image produced. So to reduce these problem
different researchers has proposed different solutions, but the best solution is to use bilateral filter. The main objective
of this paper is to find the gaps in existing literature. The different segmentation techniques are reviewed and found that
marker based is best in most of cases because it marks the regions then segment them. But optimizing the marking
regions is still an area of research.
Keywords: External markers, Gradient, Internal markers, Image segmentation, Watershed.
I. INTRODUCTION
The image segmentation is significance problem in
different fields of computer vision and image processing.
Image segmentation is the process of partitioning a digital
image into multiple segments knows as set of pixels. The
goal of segmentation is to simplify change the
representation of an image into something that is more
meaningful and easier to analyze. Image segmentation is
typically used to locate objects and boundaries (lines,
curves, etc.) in images. It is the process of assigning a
label to every pixel in an image such that pixels with the
same label share certain visual characteristics.
C. Histogram based method
Histogram based methods are very efficient when
compared to other image segmentation methods because
they typically require only one pass through the pixels. In
this technique, a histogram is computed from all of the
pixels in the image, and the peaks and valleys in the
histogram are used to locate the clusters in the image.
D. Edge detection
The first region-growing method was the seeded region
growing method. This method takes a set of seeds as input
along with the image. The seeds mark each of the objects
The result of image segmentation is a set of segments that to be segmented. The regions are iteratively grown by
collectively cover the entire image, or a set of contours comparing all unallocated neighbouring pixels to the
extracted from the image. Each of the pixels in a region regions.
are similar with respect to some characteristic or computed
property such as color, intensity, or texture. The goal of E. Watershed transformation:
segmentation operation is to simplify the image without The watershed transform has interesting properties that
make it useful for much different image segmentation
discarding important image features.
application. In image segmentation the objects in the
Image segmentation process in three stages. The first is image are separated and labelled In image segmentation
image pre-processing, then input image is converted into the objects in the image are separated and labelled for
gradient image further it is applied by watershed further analysis. This is usually done with the help of a
transform,object discrimination, where objects are grossly wide range of image segmentation techniques. Each
separated into groups with similar attributes. Third stage is technique has its own advantages and disadvantages. The
object boundary clean up, where object boundaries are effectiveness of a particular image segmentation algorithm
reduced to single-pixel widths. In recent years several is determined with respect to a particular class of images.
popular methods have been developed for image Generally a combination of two or more techniques is used
to get the desired output for a particular application. The
segmentation.[1].
scale of segmentation is application specific and even the
objects that are to be segmented. It is important to note
A. Thresholding method
The simplest method of image segmentation is called the that:
thresholding method. This method is based on a threshold 1. There is no universally applicable segmentation
value to turn a gray-scale image into a binary image.
technique that is guaranteed to work on all images.[2]
2. No segmentation technique is perfect.
B. Split and merge method
This method is used for the division of the image. If it is The goal however is to come up with an ad-hoc technique
found non-uniform (not homogeneous), then it is split into for segmentation of high resolution aerial imagery which
four son-squares (the splitting process).
is a step further in image analysis. In recent years various
Copyright to IJARCCE
DOI 10.17148/IJARCCE.2016.51243
196
IJARCCE
ISSN (Online) 2278-1021
ISSN (Print) 2319 5940
International Journal of Advanced Research in Computer and Communication Engineering
ISO 3297:2007 Certified
Vol. 5, Issue 12, December 2016
techniques have been developed for image segmentation.
One of them is watershed technique applied on images but
it has various drawbacks that is over segmentation and
sensitivity to noise so to overcome this we will use marker
based watershed technique to overcome these problems.
This technique “mark,” foreground objects and
background locations. The results obtained by the
segmentation of the image are generally very subjective,
as it depends on the information content of the image
itself. The parameters to be considered for the evaluation
of these results vary from image to image.
Markerbased watershed transform
The advantages of the watershed transformation are that it
is simple, instinctive knowledge, and can be parallelized.
The main drawback of this method is the oversegmentation due to the presence of many local minima.
To decrease the effect of severe over-segmentation,
marker-controlled watershed transformations have been
proposed. These are robust and flexible methods for
segmenting objects with closed contours. The internal
marker and external marker are initially defined. The
boundaries, even if not clearly defined, are expressed as
ridges between two markers and located. In markercontrolled watershed method to segment the image the
external marker is obtained manually by drawing a circle
enclosing object of our interest. The internal marker is
determined automatically by combining techniques
including Canny edge detection, thresholding and
morphological operation. Modify the segmentation
function so that it only has minima at the foreground and
background markers. The Watershed Transform
effectively combines elements from both the discontinuity
and similarity based methods. Since its original
development with grey-scale images, the Watershed
Transform has been extended to a computationally
efficient form (using FIFO queues) and applied to colour
images. The main advantages of the Watershed method
over other previously developed segmentation methods
are[3]
A. The resulting boundaries form closed and connected
regions. Traditional edge based techniques most often
form disconnected boundaries that need postprocessing to produce closed regions.
B. The boundaries of the resulting regions always
correspond to contours which appear in the image
obvious contours of objects.[9]
II. PROPOSED METHODOLOGY
This paper has presented an literature survey on marker
based watershed segmentation algorithms. The main
objective of this paper is to find the gaps in existing
literature. The different segmentation techniques are
reviewed and found that marker based is best in most of
cases because it marks the regions then segment them. But
optimizing the marking regions is still an area of research.
However it has found that most of existing techniques uses
Copyright to IJARCCE
the bilateral filter. But as it is known in prior bilateral filter
is unable to reduce salt and pepper noise. So in near future
to reduce this problem we will integrate marker based
watershed segmentation algorithm with hybrid median
filter to improve the performance .The main drawback of
this method is the over-segmentation due to the presence
of many local minima. To decrease the effect of severe
over-segmentation, marker-controlled by the watershed
transformations have been proposed. These are robust and
flexible methods for segmenting objects with closed
contours. The internal marker and external marker are
initially defined. The [5] boundaries, even if not clearly
defined, are expressed as ridges between marker and
locate. In marker-controlled watershed method to segment
the image the external marker is obtained manually by
drawing a circle enclosing object of our interest. A gray level co-occurrence matrix. Then, both gradient images are
fused to give the final gradient image. After the initial
results of segmentation, we use the merging region
technique to remove small regions. It is well known that
the main problem of this method is that the images we
consider are often noisy, which implies that we have a lot
of local minima and this leads to an over segmentation.
We propose in this paper, a new method for decreasing the
over segmentation of standard watershed based
techniques. Our method is based on the topological
gradient approach. The topological gradient has here the
interesting property to give more weight to the main
edges, it provides a more global analysis of the image than
the Euclidean gradient or the image[10]. to noise as we
show it in the numerical applications section. The gradient
image is often used in the watershed transformation,
because the main criterion of the segmentation is the
homogeneity of the grey values of the objects present in
the image. But, when other criteria are relevant, other
functions can be used. In particular, when the
segmentation is based on the shape of the objects, the
distance function is very helpful.[6].
III. RESULTS AND DISCUSSIONS
We consider in this section the problem of denoising of an
image and preserving features suchas edges. According to
the previous section, thetopological asymptotic analysis
provides the location of the edges as they are precisely
defined as the most negative points of the topological
gradient,the results obtained by the topological gradient
algorithm. The image processed given by a 256×256 gray
level image and represents some rice grains. Fig. 1b is
obtained by adding to the original image a Gaussian noise
( = 20). The reconstructed image is shown in Fig. 1c:
the topological gradient method for the restoration process
was applied with c0 = 1. Finally, we give in, the edges of
the reconstructed image. To obtain the restored image, the
topological gradient algorithm requires only 3 system
resolutions for calculating u0, v0 and the restored image
u1 given respectively by Eqs. 4, 6, and 2. For a better edge
preservation, one has to threshold the topological gradient
with a small enough coefficient. In the other case, if the
DOI 10.17148/IJARCCE.2016.51243
197
ISSN (Online) 2278-1021
ISSN (Print) 2319 5940
IJARCCE
International Journal of Advanced Research in Computer and Communication Engineering
ISO 3297:2007 Certified
Vol. 5, Issue 12, December 2016
thresholding coefficient is set to a large value, then the
edges obtained will be thick, leading to an over smoothing
and a loss of an important edge information and then a
degradation of the restored image. Finally, to speed up the
computations, a spectral method based on the discrete
cosine transform has been used for the resolution of the
direct and adjoint problems. Since the coefficient c is
equal to a constant c0 except on edges, then the discrete
cosine transform is a good preconditioner for the
conjugate gradient method. The complexity of the
restoration algorithm is O(NlogN) where N is the number
of pixels of the image. Some comparisons about the
computationtimes with other classical methods are
presented in Jaafar Belaid et al. (2008).
Segmentation using a classical watershed
The goal of this section is to present numerical tests for the
segmentation problem using mathematical morphology
tools. The approach used in this work is based on the
watershed transform. One should remark that we can
either define the watershed of the function u or of its
gradient: the difference between the two definitions is that
in the first case we obtain the influence zones of the
processed image, while the second case gives the image
edges. In both cases, the watershed gives an
oversegmentation and to avoid this drawback, a markers
technique can beused. We propose in this section to give
numerical results based on this classical method according
to the work of Beucher (1990), in which the author has
proposed to use both influence zones and minima ofthe
filtered image as marker criteria. Fig. 2 illustrates this
approach. We have considered the same original image as
previously shows the watershed of the image shows the
watershed of its gradient. The over segmentation is clearly
seen. We mention here, that the numerical tests given by
give a segmented image with 1905 regions for a
computational time of 550 s. This over segmentation can
be in a first step corrected by applying a morphological
filter shows the watershed of the filtered image: the
number of regions segmented is attenuated (868 regions)
for a computational time of 775 s, but the segmentation [7]
result remains unacceptable.
However, as over segmentation is due to the fact that we
obtain a lot of minima, and the use of morphological filters
can only suppress some of them, then another way to act
on these minima is to apply the swamping approach, by
imposing markers for new minima.
The original image is the input image here we are using
the lena image as the input image .Then the further preprocessed by using gradient magnitude as the
segmentation function. The watershed transform cannot
be applied directly to the binary image hence it is
converted into RGB for better pixel response.
Fig3:Gradient magnitude of the image
Fig.4 Thresholder image
Fig. 2: Original image
Copyright to IJARCCE
Fig.5 .Segmented image
DOI 10.17148/IJARCCE.2016.51243
198
IJARCCE
ISSN (Online) 2278-1021
ISSN (Print) 2319 5940
International Journal of Advanced Research in Computer and Communication Engineering
ISO 3297:2007 Certified
Vol. 5, Issue 12, December 2016
IV. CONCLUSION
This paper has presented an literature survey on marker
based watershed segmentation algorithms. The main
objective of this paper is to find the gaps in existing
literature. The different segmentation techniques are
reviewed and found that marker based is best in most of
cases because it marks the regions then segment them. But
optimizing the marking regions is still an area of research.
However it has found that most of existing techniques uses
the bilateral filter. But as it is known in prior bilateral filter
is unable to reduce salt and pepper noise. So in near future
to reduce this problem we will integrate marker based
watershed segmentation algorithm with hybrid median
filter to improve the performance of proposed
algorithm.[11]
REFERENCES
Baojingji,jianping Lv and CaiXia Zhoa, ”improved watershed
algorithm based on segmentation “,Xian institute of posts and
telecommunications pp.103-107.
[2] Boren Li and Mao Pan, “An Improved
Segmentation of High
Spatial Resolution Remote Sensing Image using Marker-based
Watershed Algorithm,‟‟ pp. 98-104, IEEE,2012.
[3] Chen Wei-bin and Wenzhou Zhejiang, “A
New Watershed
Algorithm for Cellular Image Segmentation Based on Mathematical
Morphology,” International Conference on Machine Vision and
Human-machine Interface, vol.53, pp.2405-2414, IEEE, 2010.
[4] GuiMei Zhang and Ming-Ming Zhou, “Labelling watershed
algorithm based on Morphological Reconstruction in Color Space,”
pp.51-55, IEEE, 2011.
[5] Jun Tang, “A Color Image Segmentation algorithm Based on
Region Growing,” vol.6, pp.634-637, IEEE, 2010.
[6] Li Cheng, Li Yan and Fan Yan Shangchun,“CCD infrared image
segmentation using
watershed algorithm,” Third International
Conference on Measuring Technology and Mechatronics
Automation, vol.1, pp.680-683, IEEE,2011.
[7] Shiping Zhu, Xi Xia and Qingrong Zhang, “An Image
Segmentation Algorithm in Image Processing Based on Threshold
Segmentation,” Third International IEEE Conference on SignalImage Technologies and internet based system, IEEE, 2008.
[8] Quan Longzhe, “Automatic Segmentation
method of Touching
Corn Kernels in Digital Image Based on Improved Watershed
Algorithm,” pp.34-37, IEEE, 2011.
[9] Wei Zhang, “The Marker-Based Watershed Segmentation
Algorithm of Ore Image,” pp.472-474, IEEE, 2011.
[10] Xiaoyan Zhang, Lichao Chen, Lihu Pan and Lizhi Xiong, “Study
on the Image Segmentation Based On ICA and Watershed
Algorithm,” Fifth International Conference on Intelligent
Computation Technology and Automation, pp. 978-912, IEEE
2012.
[1]
Copyright to IJARCCE
DOI 10.17148/IJARCCE.2016.51243
199