NOVATEUR PUBLICATIONS

International Journal of Research Publications in Engineering and Technology [IJRPET]

ISSN: 2454-7875
VOLUME 3, ISSUE 5,May -2017
COLOR IMAGE SEGMENTATION TECHNIQUE USING COOPERATIVE
BACTERIAL FORAGING ALGORITHM.
AMIT D.PUROHIT
BNCOE, Pusad (India), [email protected]

PROF. S. T. KHANDARE
BNCOE, Pusad (India), [email protected]

ABSTRACT: The basic method of image segmentation is

Image segmentation is a crucial and thresholding. From a grayscale image, thresholding can be
challenging problem in image processing and often a used to generate binary images. The idea use in this
basic step for high level analysis. The intent of image method is to select the threshold value. Lots of
segmentation is to divide an image into different conventional methods are used in engineering including
classes based on features, such as color, intensity or the k-means clustering, Otsu's method that uses maximum
histogram, where each pixel in the image should go to entropy and maximum variancemethod. The main intent is
one class and only one class. According to the that, the proposed segmentation can be work effectively
thresholds the segmented results whether or not for image, based on automatic thresholding and color
consistent to the image is also an issue should be model based image segmentation.Image segmentation is
considered.Here wepropose a new method for color considered as an important basic operation for
image segmentation using multilevel thresholding. interpretation of acquired images and for meaningful
This paper proposes multilevel thresholding for color study or analysis of the images. It is a classic inverse
image segmentation using Cooperative Bacterial problem which consists of achieving a compact region-
Foraging Algorithm (CBFA), which mixedcell-to-cell based description of the image scene by decomposing it
communication, bacterial chemo taxis, and an into meaningful or spatially coherent regions sharing
adaptive foraging mechanism to enhance original similar attributes.
Bacterial Foraging Algorithm (BFA) which is One of the key problems in color image analysis is
motivated by the social foraging behavior of that of segmentation. Image segmentation is a technique
Escherichia coli(E. Coli Bacteria). and process which divide the image into different feature
of region and extract out the interested target. Properties
I.INTRODUCTION: like intensity, texture, depth, gray-level, color help to
Image segmentation is a technique to divide an recognize similar regions; such properties are usedto form
image into multiple segments and extract the meaningful groups of regions having a similar meaning. Segmentation
objects, which is the difficult step in image processing and is a valuable tool in several fields including industry,
image analysis. The aim of dividing an image into different health care,pattern recognition, image processing, content
regionis to make an image more meaningful and easier to based image, remote sensing, traffic image, videos and
analyze and understand. Here and now image computer vision.After complete study of different
segmentation has been universally used in many practical methodshere we will projected the new thresholding
applications such as medical imaging, remote sensing, mechanism. To achieve thresholding, global and local
Optical Character Recognition(OCR) and object detection. analysis, and assessment of the color image data will be
The intention of segmentation is to clarify the explored. Main aim of our work is to develop a
image and change the representation of an image into segmentation mechanism for color images based on
something that is moreimportant, useful and easier to multilevel thresholding. To accomplish the main goal of
study or examine. Image segmentation is mostly used to the work focus is on automatic thresholding by
detect objects and background in images. More precisely, representing image into completely different color models
image segmentation is the mechanism of assigning a label [Yuan and Chen, 2009].
to each pixel in an image such that pixels with the same
tag share certain visual characteristics. Image III.PROPOSED SYSTEM DESIGN
segmentation is an important signal processing tool that is Here weproposecolor image segmentation
universally occupied in many applications including technique using CBFA. Here we extend original Bacterial
medical imaging, remote sensing, Optical Character Foraging Algorithm (BFA)and invent Cooperative Bacterial
Recognition (OCR) and object detection. Foraging Algorithm (CBFA) for the segmentation of an
image, wecombined cell-to-cell communication,bacterial

56 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 5,May -2017
chemo taxis, and an adaptive foraging mechanism. The FOUR MAIN STEPS IN CBFA:
suggested algorithm can take less CPU processing time,  CHEMOTAXIS:
gives us superior quality more convenient segmentation, This procedure reenacts the development of an
which is proved by comparing CBFA with other algorithm. E.coli cell (Bacteria) through swimming and tumbling
In this paper, the innovative Cooperative Bacterial by means of flagella. Naturally an E.coli bacterium can
ForagingAlgorithm is applied to find the optimal move in two diverse ways. It can swim for a timeframe
thresholdvalues by applying two manipulated steps, in a similar heading(direction) or it might tumble (by
namely acell-to-cell communication and a self-adaptive and large unique bearing), and exchange between
foragingstrategy, which is extended by the classical BFA these two methods of operation for the whole
algorithm. lifetime[10].
The proposed method which is proved that the
Multi-levelthresholding method based on Cooperative  SWARMING:
Bacterial ForagingAlgorithm (CBFA) forcolor image It is always wanted that the bacterium that has
segmentation is considered as an optimizationproblemcan searched the most advantageous path of food should
execute faster than the othertraditional algorithmsand is try to attract other bacteria so they reach the desired
more stable. place more quickly.Swarming makes the bacteria
bunch up into groups and hence move as a concentric
1. COLOR IMAGE SEGMENTATION STEPS: pattern of groups with high bacterial density [10].
A] THE CONVERSION OF COLOR SPACE:
Here RGB components of color image are  REPRODUCTION:
converted to HSV vector. The conversion formula of RGB The least healthy bacteria eventually die while
and HSV is shown in Eq. (1): each of the healthier bacteria (thoseyielding lower
value of the objective function) asexually divided into
2 bacteria, which are thenplaced in the same location.
This keeps the swarm size as it is[10].

 ELIMINATION/ DISPERSAL:
Quick changes in the local environment where a
bacterium community lives may occur due to various
B] MULTILEVEL THRESHOLDING FOR IMAGE reasons e.g. a significant local rise of temperature may
SEGMENTATION: kill a group of bacteria that are currently in a region
We get three separate vectors of HSV for solving the with a high concentration of nutrient gradients.
problem for multilevel thresholdvaluesand then apply Events can take place in such a fashion that all the
CBFA (Cooperative bacterial foraging algorithm) to each bacteria in a region are killedor a collection is
vector to get the corresponding segmentation image at the dispersed into a brand new location [10].
three vectors.Combine the three vector results, and then We are testedthe performance ofthe proposed
get the final segmentation image. algorithmwith different standardtest images and
compared the result with traditional Bacterial
ForgingAlgorithm.According to results the Cooperative
Bacterial ForagingAlgorithm is better than the other
algorithms, not only in termsof solution quality, but also in
terms computational efficiencyandstability, especially
when the image segmentation of themulti-level is
processed, the inventedtechnique for imagesegmentation
based on the Cooperative Bacterial ForagingAlgorithm
(CBFA)gives usimproved performance to find the
superiorthresholds in less CPU processing timewith more
stability.

Fig: Flow Chart for CBFA

57 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 5,May -2017
Standard deviation and CPU Processing Time:
In statistic, the standard deviation(STD, also
represented by the Greek letter sigma σ) is a measure that
is used to quantify the amount of variation or dispersion of
a set of data values. A low standard deviation indicates
that the data points tend to be close to the mean (also
called the expected value) of the set, while a high standard
deviation indicates that the data points are spread out
over a wider range of values.We are going compare BFA
and CBFA according to STD(Standard Deviation) and CPU
Processing Time. According to results CBFA is more stable
than BFA because CBFA has less STD (Standard Deviation)
Fig: BFA ofLighthouse, (a) Original (b)(c)(d) The result of
segmentation with 2, 3, 4thresholds. values. Also it required less CPU processing time for image
to find multi-level thresholds in comparison to the
(Bacterial Foraging Algorithm) BFA.
The given image contains multiple features. The
pixels of the image would be divided into n classes. In
order to judge the stability of the algorithm, the standard
deviation can be calculated by the STD, which is defined
as:

Assumingσias the best fitness value of the ith runs

Fig:CBFAofLighthouse, (a) Original (b)(c)(d) The result of of thealgorithm, µ as the average value of σiand N as
segmentation with 2, 3, 4 thresholds. therepeated times of each algorithm

Table: STD OF DIFFERENT METHODS

Test image Thresholds STD of STD of
BFA CBFA
2 0.3676 0.3567
3 0.3742 0.3424
Lighthouse 4 0.3269 0.3236
2 0.3597 0.3580
3 0.3479 0.3466
Tulips 4 0.3429 0.3369

Fig: BFA ofTulips, (a) Original (b)(c)(d) The result of Table: CPU PROCESS TIME OF DIFFERENT METHODS
segmentation with 2, 3, 4 thresholds. Test Thresh Computatio Computatio
image olds nal nal
time of BFA time of
CBFA
2 2.7260 0.9365
3 4.6353 3.5999
Lighthou 4 9.1040 8.8013
se
2 1.0519 0.9103
3 3.9908 2.0551
Tulips 4 6.7200 5.6430
Fig: CBFA ofTulips, (a) Original (b)(c)(d) The result of
segmentation with 2, 3, 4 thresholds.

58 | P a g e
 NOVATEUR PUBLICATIONS
International Journal of Research Publications in Engineering and Technology [IJRPET]
ISSN: 2454-7875
VOLUME 3, ISSUE 5,May -2017
Our proposed system is more useful to solve the Journal of Electronic Imaging, vol.13(1), pp.146–168.
problem as compare to conventional Bacterial Foraging 2004.
Algorithm such as BFA and this method is more suitable 7) [Rafika HARRABI and Ezzedine BEN BRAIEK
when the image segmentation of multilevel is under 2014]Color Image Segmentation by Multilevel
consideration, which can find the better thresholds in less Thresholding using a Two Stage Optimization Approach
CPU processing time with more stability. and Fusion (IJEIT) Volume 3, Issue 11.
8) [S. Dasgupta, S. Das, A. Abraham, A. Biswas 2009]
IV.CONCLUSION Bacterial Foraging Optimization Algorithm Theoretical
The thresholding method is universally used for Foundations, Analysis, and Applications.
the segmentation of images due to its accuracy, simplicity,
and robustness. Efficiently selecting the optimal threshold
to group pixels within meaningful regions is the key of the
thresholding method.Here we propose a new color image
segmentation technique, based Cooperative Bacterial
Foraging Algorithm which aim to use maximum number of
features of the same color image in order get a more
valuable and accurate segmented effect.Here useful
conversion of color spaces is done. Here we perform
feature(RGB component) extraction and color space
conversion to HSV. Then we make use of the CBFA
algorithm which determine threshold automatically from
the picture content and perform the segmentation.
According to results CBFA is more stable than BFA
because CBFA has less STD values. Also it required less
CPU processing time for image to find multi-level
thresholds in comparison to the BFA method.

REFERENCES
1) [Yuan and Chen, 2009] Image Segmentation Method
Using Thresholds Automatically Determined from
Picture Contents, EURASIP Journal on Image and Video
Processing Volume 2009, Article ID 140492, 15 pages
doi:10.1155/2009/140492.
2) [S. Dasgupta, S. Das, A. Abraham, A. Biswas] Adaptive
computational chemo taxis in bacterial foraging
optimization: an analysis, IEEE Transactions on
Evolutionary Computing, vol.13 (4), pp.919–941.
2009.
3) [Harrabi and Braiek, 2012] Color image segmentation
using multi-level thresholding approach and data
fusion techniques: application in the breast cancer cells
images, EURASIP Journal on Image and
VideoProcessing.
4) Chen, H. N., Zhu, Y. L., Hu, K. Y., “Multi-colony bacteria
foraging optimization with cell-to-cell communication
for RFID network planning�, Applied Soft Computing,
vol.10 (2), pp. 529-551, 2010.
5) Liu,Y, Tian,L.W., Hu,K.Y., Zhu, Y. L., Computational
foraging in bacterial colony over composition
environments, Journal of Pure and Applied
Microbiology, vol. 7(2), 2013.
6) Sezgin, M.,&Sankur, B. Survey over image thresholding
techniques and quantitative performance evaluation.

59 | P a g e