Professional Documents
Culture Documents
ABSTRACT
Image segmentation is the process of partitioning an
image into multiple segments, so as to change the
representation of an image into something that is more
meaningful and easier to analyze. Several generalpurpose algorithms and techniques have been developed
for image segmentation. Image segmentation refers to
partition of an image into different regions that are
homogenous or similar and inhomogeneous in some
characteristics as the new graph cut approach is an
emerging technique for image segmentation, as it can
minimize an energy function composed of data term
estimated in feature space and smoothness term
estimated in an image domain. Previous approaches
using graph cuts have shown good performance for
image segmentation, they manually obtained prior
information to estimate the data term, thus automatic
image segmentation is one of issues in application using
the graph cuts method. As it requires low computational
complexity and is therefore very feasible for real-time
image segmentation processing. In the experiments, we
investigated problems of previous methods such as
mean shift segmentation, watershed technique and
automatic graph cut based image segmentation As a
result, the graph cut method showed better performance
than previous methods.
Keywords: - Mean Shift, Watershed, Graph Cuts, Edge
Based Segmentation.
1. INTRODUCTION
Image processing refers to processing of a 2D picture
by a computer. An image is defined in the real world
is considered to be a function of two real variables. An
image is an array, or a matrix, of square pixels (picture
elements) arranged in columns and rows, which
identifies a point in the image and the corresponding
matrix element value identifies the gray level at that
point. Image processing is a method to convert an image
into digital form and perform some operations on it, in
order to get an enhanced image or to extract some
useful information from its image domain [1].Image
Segmentation is the process of partitioning a digital
image into multiple regions or sets of pixels. Actually,
partitions are different objects in image which have the
same texture or color. All of the pixels collectively
cover the entire image, or a set of contours to some, or
www.ijsret.org
682
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 3, June 2014
2. PREVIOUS APPROACHES
Mean shift method
Its a non-parametric feature-space analysis technique, a
so-called mode seeking algorithm. Mean shift is a
procedure for locating the maxima of a density function
given discrete data sampled from that function. It is
useful for detecting the modes of this density. This is an
iterative method, and we start with an initial estimate x.
Let a kernel function ( )be given. This function
determines the weight of nearby points for re-estimation
of the mean [10]. Typically Gaussian kernel on the
distance to the current estimate is used, ( ) =
( )
( )
(1)
www.ijsret.org
683
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 3, June 2014
{ , }
).
)
(2)
In the energy minimization equation, the first term
called as data term consists of constraints from the
observed data and measures how the labels are
assigned. Label fp fits with site p and is measured by
Dp. The second term which is the smoothness term
measures to what extent f is not piecewise smooth. N
represents the neighborhood system like 4 or 8connected system. If f = f , T(f f )becomes 0 and
1 otherwise. In image segmentation it is expected the
boundary to be positioned on the edges. Hence the
typical selection of
is:
=
1
( , )
(3)
+
Optimization: The segmentation can be estimated as a
( , , , ) [7].
global minimum over :
www.ijsret.org
684
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 3, June 2014
4. EXPERIMENTS RESULTS
To verify the effectiveness of the proposed graph cut
method, quantitative evaluation was tested with natural
color and the real MRI brain images, and for
comparison, two methods were applied to the same test.
The performance of image segmentation on the test set
was evaluated in terms of visually and the subjective
parameters such as PSNR, MSE and the computational
TIME.
Table 1 shows list the PSNR, MSE and TIME on the
test set. The proposed method outperformed the
previous two methods on the test set, and it was closer
to human segmentation than other algorithms. The
problems of previous two methods for image
segmentation are as follows the result of The MS
undesirably produced a very large number of small but
quasi-homogeneous regions as shown in figure shown.
This is equal to the problem of the image domain-based
segmentation, as the mean shift procedure is performed
only in an image domain. Thus, the MS is sensitive to
noises, as affected by edge information. Therefore, the
gradient based method (watershed technique) is also not
found suitable as these are observed that they are highly
sensitive to noise and not segmenting the discontinuities
in the image, only identifies the outer boundaries so this
is a major drawback of the watershed algorithm.
Image
3
Graph cut
(k-nearest)
Mean shift
Watershed
Graph cut
(adjacant)
Graph cut
(k-nearest)
28
12
27
28
100
400
0
127
109
3.77
4.32
1.55
1.42
29.5
170
1.5
www.ijsret.org
685
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 3, June 2014
80
60
PSNR 40
20
0
image 3 image 4 image 5
MEAN SHIFT
WATERSHED
GRAPH CUT(A)
GRAPH CUT(K)
MEAN SHIFT
WATERSHED
GRAPH CUT(A)
GRAPH CUT(K)
MEAN SHIFT
WATERSHED
GRAPH CUT(A)
GRAPH CUT(K)
www.ijsret.org
686
International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 0882
Volume 3, Issue 3, June 2014
5. CONCLUSION
From the results obtained, it is evident that the
automatic graph cut method is capable of accurately
segmenting any of the arbitrary shapes at the cost of
computational time, as it overcame the problems from
the graph cut using GMM approach of segmentation
which is only applicable to the hyper-spherical and
hyper-ellipsoidal classes. This technique is also highly
efficient in segmenting the brain tissues and other brain
element in brain MRI images.
REFRENCES
[1] H.P. Narkhede, Review of Image Segmentation
Techniques International Journal of Science and
Modern Engineering (IJISME) ISSN: 2319-6386, Vol.1,
Issue 8, July, 2013.
[2] S. Tripathi, K. Kumar, B.K.Singh, R.P.Singh,
Image Segmentation: A Review, International Journal
of Computer Science and Management Research, Vol.
1, Issue 4, November, 2012.
[3] L. Lucchese and S.K. Mitra, Color Image
Segmentation: A State-of-the-Art Survey, Proceedings
of the Indian National Science Academy (INSA-A),
67(2): pp. 207221, 2001.
[4] L. Vincent and P. Soille, Watersheds in Digital
Space: An Efficient Algorithm based on Immersion
Simulation, IEEE Transactions on Pattern Analysis
and Machine Intelligence, 13(6): pp. 583597, 1991.
[5] D. Comaniciu and P. Meer, Mean Shift: A Robust
Approach toward Feature Space Analysis, IEEE
Transactions on Pattern Analysis and Machine
Intelligence, 23(5): pp. 603619, 2002.
[6] S. Makrogiannis, G. Economou, and S. Fotopoulos,
A Region Dissimilarity Relation that Combines
Feature-Space and Spatial Information for Color
Segmentation, IEEE Transactions on System, Man,
and Cybernetics, Part B, 35(1): pp. 4453, 2005.
[7] A. Park, J. Kim, S. Min, S. Yun, K. Jung, Graph
Cuts based Automatic Color Image Segmentation using
Mean Shift Analysis IEEE Digital Image Computing:
Techniques and Applications, 2008.
[8] Y. Boykov, V. Kolmogorov, An experimental
comparison of min-cut/max-flow algorithms for energy
minimization in vision, IEEE Trans, Pattern Anal.
Machine Intell., 26, pp. 11241137, 2004.
[9] Geman, S., Geman, D., Stochastic relaxation, gibbs
distributions, and the Bayesian restoration of images,
IEEE Trans, Pattern Anal. Machine Intell, pp. 721741,
1984.
[10]M. Sonka, V. Hlavac and R. Boyle, Image
processing, analysis, and machine vision, Third
edition, Thomson, USA, 2008.
www.ijsret.org
687