International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.

AN INTEGRATION OF GENETIC ALGORITHM AND PROJECTED CLUSTERING FOR
OPTIMIZATION OF CONTENT BASED IMAGE RETRIEVAL SYSTEM
S.SELVAM M.Sc., M.Phil.,
Dr.S.THABASUKANNAN MCA.,
Research Scholar,
M.Phil., MBA., M.Phil.,M.Tech., Ph.D,
Research and Development Center,
Principal, Pannai College of Engg&Tech
Bharathiar University,
Sivagangai 630 561, Tamilnadu, India
Coimbatore-641046.
E-Mail: thabasukannan@gmail.com,
Assistant Professor,
Mobile: (91)9944436258, 7200467890.
Department of Computer Applications,
N.M.S.S.Vellaichamy Nadar College, Madurai-19. .
E-Mail:s.selvammscmphil@gmail.com
Mobile :( 91)9894226757
Abstract:
Inrecentyearsespeciallyinthelastdecade,therapi
digitization, the volume of digital data is
ddevelopmentincomputers,storagemediaanddi

increasing tremendously. During retrieval of

gitalimagecapturingdevicesenabletocollectalar

images from computer, some problems crop

genumberofdigitalinformationandstorethe

up. To solve this problem we use CBIR.

minicomputerreadableformats.Thelargenumbe

Toretrieveanyimage,wehavetosearchforitamo

rsofimageshasposedincreasingchallengestocom

ngthedatabaseusingsomesearchengine.Then,t

putersystemstostoreandmanagedataeffectivelya

hissearchenginewillretrievemanyofimagesrela

ndefficiently.Althoughthisareahasbeenexplore

tedtothesearchdone.Themainproblemfo r t he

dfordecadesandmanyresearchershavebeendone

user

todevelopsomealgorithmsthatsolvesomeofitspr

isthedifficultyoflocatinghisrelevantimageinth

oblems,notechniquehasachievedtheaccuracyof

islargeandvariedcollectionofresultedimages.T

humanvisualperceptionindistinguishingimages.

osolvethisproblem,text-basedandcontent-

To

basedarethetwotechniquesadoptedforsearcha

fulfill

the

latest

technological

requirements some domains like commerce,

ndretrieval.

academic, health care, police etc., use

The

information in the form of images.By using

tobuildmoregeneralizedCBIRsystemwhichincr
8

main

objective

of

this

paper

is

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
easethesearching abilityand
accurateresults.

provide

more

optimum manner. For this some data mining

Toimprovetheretrieval

approach

and

heuristic

approach

are

accuracy

needed.The similarity checking algorithm is

thesystemhastakenthefeedbackfromtheuser

surely increases the efficiency in terms of

automatically.

retrieval.

To evaluate the performance of new system,

Nowadays

we useWANGdatabase. The metrics used for

archive has a size of Petabytes (PB) and grows

evaluation

to Terabytes (TB). It triggers the demand of

are

precision,recallandretrievaltime.

The

qualitative

national

and

geographic

quantitative

imagery

image

performance can be evaluated by comparing

retrievalsystems. An image retrieval system is

some existing systems in CBIR. The

a computer based system for browsing,

performance of new system in terms of the

searching andretrieving images from a large

metricsproves togood.

database of digital images. Searching and

Keywords:
CBIR, Genetic Algorithm, HARP Algorithm,
Precision, Recall.

retrieving is not bit bybit comparison. It is not

1. Introduction

The drawbacks of the TBIR initiate to do the

To meet the interested requirements of our

research in the field ofCBIR. In CBIRalso

collection of imagesfor future research, we

known

have

image

retrieval is based on the image contents. Many

libraries.To do the above in an effective

techniques have been developedfor the most

manner some domain independent algorithms

important CBIR systems, which is a system,in

are required. Many algorithms are needed to

which retrieves visual-similar images from

represent, describe and retrieve images in an

largeimage

to

search

various

built-in

a matching process on the raw data.

as

querybyimagecontent

database

based

(QBIC),

on

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
automaticallyderivedimage

which

between twoimages, the simplest way is to

has been a very active area recently. In most

calculate the distance between the feature

of the existing CBIR systems,the image

vectors representingthe two images. To find

content

low-

more similar or relative images, the heuristic

levelfeatures such as colour, texture and

approach based Geneticalgorithm has been

shape. Thedrawback of low-level features is

used in the CBIR system.

losing much detailinformation of the images,

Image retrieval techniques are useful in many

in case of looking for imagesthat contain the

image-processing applications. Content-based

same object or same scene withdifferent

image retrieval systems work with whole

viewpoints. In recent years, the interest

images and searching is based on comparison

pointdetectors and descriptors are employed in

of the query. General techniques for image

manyCBIR systems to overcome the above

retrieval are color, texture and shape. These

drawback.

techniques are applied to get an image from

Similarity can be defined as the quantitative

the image database. They are not concerned

measurement that indicates the strength of

with the various resolutions of the images,

relationship between two image objects.

size and spatial color distribution. The content

Dissimilarity

and metadata based system gives images using

is

represented

is

also

features,

by

their

quantitative

measurement that reflects the discrepancy

an effective image retrieval technique.

between two image objects.

The main aim of this new system is to

In a CBIR system, the retrieval of images has

minimize the computation time and user

been done by similarity comparison between

interaction. In conventional CBIR, the time

thequery image and all candidate images in

taken to analyze the output images is more

the database. To evaluate the similarity

because the output displays at the end of the

10

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
process. But in our newly constructed system,

Clustering

the time taken for analysis is meager, because

patternsintogroups.Itsmaintaskistoassigningase

it passes through various processing stages

tofobjectsintogroupssothat

based on the users threshold values.The step

theobjectsinthesameclusteraremoresimilartoea

of this study is to reduce the gap between high

chotherthantothoseinotherclusters.

and low level features as CBIR calculates the

Inthis

similarity between user query and repositories

paper,Color,TextureandShapefeatureswereex

image. It may lead to unwanted retrieval of

tractedandcombinedtoformfeaturevectorofim

images. By using HARP, it groups the output

age.Forcolorfeatures,themomentsofthecolordi

images and a representative image from every

stributionwerecalculatedfromtheimagesandus

cluster.

edascolordescriptor.Fortexturefeatures,

The

subsequent

process

is

istheunsupervisedclassificationof

performance evaluation on the basis of speed

weusedGaborfilter,whichisapowerfultexturee

and accuracy, because it gives strong impact

xtractiontechniqueindescribingthecontentofi

on the implementation.

mage.Forshapefeatures,edgehistogramfeature

InsteadofRelevance

sthatincludefivecategorieswereusedasshapede

Feedbackwecanuseanyclusteringalgorithmtha

scriptor.Thesethreedescriptorswerecombined

tbasedonthefeaturesextractedfromtheimagest

andoptimizedusingGAwith

hemselves,andallocatesthoseimagesintothene

clu st ering

arestcluster.Thealgorithmcalculatesandallocat

accuracyasafitnessfunctiontoselectoptimumw

esuntilthereislittle variation in the movement

eightsoffeatures.WeperformedGAwithH A R

of featurepoints ineachcluster.

P
onthedatabaseasanofflinestep,

HARP

clu st er ing
and

thesystemdoesnotneedtosearchtheentiredatab

11

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
aseimages;insteadjustanumberofcandidateima

results show that the reranking algorithm

gesarerequiredtobesearchedforimage

achieves a more rational ranking of

similarity.

retrieval results and it is superior to

2. Previous Study

Reranking via partial Grouping method

c) Lijun Zhao et.al.have proposed a multi-

There are various approaches are present for

round relevance feedback (RF) strategy

CBIR. Some of the important literature which

based on both support vector machine

covers the more important CBIR System is

(SVM) and feature similarity to reduce the

discussed below.

gap between query and retrieve result.

a) Chin-Chin Lai et.al. have proposed an

d) SharadhRamaswamyet.al.have published a

interactive genetic algorithm (IGA) to

paper on a fast clustering-based indexing

reduce the gap between the retrieval

technique. In this method relevant clusters

results and the users expectation .They

are

have used color attributes like the mean

have

also

used

till

the

exact

nearest

neighbors are found. This enables efficient

value, standard deviation, and image

bitmap.They

retrieved

clustering with low preprocessing storage

texture

and computation costs.

features like the entropy based on the gray

e) Nhu-Van Nguyen et.al. have proposed

level co-occurrence matrix and the edge

Clustering and Image Mining Technique

histogram.

for fast Retrieval of Images. The main

b) Zhang Xu-boet.al.have published a paper

objective of the image mining is to remove

on improved K-means clustering and

the data loss and extracting the meaningful

relevance feedback to re-rank the search

information to the human expected needs.

result in order to remedy the rank

The clustering-repeat gives good result

inversion problem in CBIR. Experimental

12

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
when the number of examples of feedback
is small.
f) Hua Yuan et.al. have presented a new
statistical

model-based image feature

3. Architecture of new CBIR system

extraction method in the wavelet domain

and a novel Kullback divergence-based
similarity measure. The Gaussian Mixture
Models(GMM) and Generalised GMM are
presented to help extract new image
features.
From the literature survey it is concluded that
a wide variety of CBIR algorithms have been
proposed in different papers. The selection
feature is one of the important aspects of
Image Retrieval System to better capture

Figure 1 A New architecture for CBIR

users intention. It will display the images

Training

from database which are the more interest to

Image

Input:

Thelearningphasetellsaboutthetrainingproc

the user.

esswhichahugeamountsampleimagesareinp
utinthefirststep. The geneticalgorithm is
used

to

trainthefeatureswithdifferentweights.For
optimizing

thefeatureweights

fitnessfunction,
13

HARP

and

algorithm

for
is

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
used.Thetrainingpartoutputstheclassifyingr

histogram,

color

momentsand

esultandstoresitinthefeaturedatabase.Allthe

coherence vectorstoring, filtering and

sestepsperformedofflineandeachclasswillb

retrieving

eindexingalongwithitsassociatedclassIDint

emerging MPEG-7 is a new multimedia

heindexfiles.

standard, which has improved CBIR by

Feature Extraction (Image signature):

providing a rich set of standardized

There are various kinds of visual features

descriptors and description schemas for

to represent an image, such as color,

describing multimedia content. MPEG-7

texture, shape, and spatial relationship.

has included dominant color, color

Since one type of features can only

structure, scalable color, and color

represent part of the image properties, a lot

layout as color features. Here we used

of work done on the combination of these

Color Structure Descriptor (CSD) as

features.

color feature.The CSD represents an

Thefeatureofeachimageisverymuchsmaller

image by both the color distribution of

insizecomparedtotheimagedata,so

the image or image region and the local

thefeaturedatabasecontainsanabstractionoft

spatial structure of the color. CSD used a

heimagesintheimagedatabase.

8 8 structure to scan the total image.

audiovisual

data.

color

The

a) Colors: are defined on a selected color

This descriptor counts the number of

space. Varieties of color spaces include,

times a particular color is contained

RGB, LAB, LUV, HSV (HSL), YCrCb

within the structuringelement while the

and the HMMD.Common color features

image or image region is scanned by this

or

structuring element.It has used HMMD

descriptors

in

CBIR

systems

includecolor-covariance matrix, color

color space.

14

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
b) Texture:There exist different approaches

Selection: Duringeachiterationexisting input

to extract and represent textures. They

iterations are refreshed through a fitness

can be classified into space-based,

process. If any iteration is best fit, then it is

frequency-based models, and texture

the solution and no further selection is

signatures. Some popular techniques i.e

needed.

wavelet

Mutation is an interchange of data from one

transform,

co-occurrence

matrix, and Gabor filters are applied to

iteration to another.

express texture features for image.

Cross over is a process n-parent solutions

c) Shape:It is seen that natural objects are

used to derive a child solution.

primarily recognized by their shape.

HARP-a

Clusteringalgorithm:

Two main types of shape feature are

algorithm

is

commonly used; global features such as

approach.Initiallypick

aspect ratio, circularity and moment

among all current cluster on the basis of

invariants and local features such as sets

smallest distance by merging all the selected

of consecutive boundary segments.

and related data on various clusters.

based

on
up

The

bottom-up

each

element

Genetic Algorithmfor optimization:It is

In HARP algorithm, the accuracy level of

used

complicated

clustering is more by using relevance

problems. It is based on heuristic approach

indexing and merge score. The scalability

that imitates the process of natural selection.

level is also very high. The time taken for

It is used exclusively for the purpose of

finding the closest cluster is very less.

optimization. Each intermediary solution can

Database: A database containing number of

be mutated and changed. It contains the

images with any one of the formats.bmp,

following operators:

.jpg, .tiff. is required.

to

find

solution

to

15

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
Query: The user provides a sample image or

4. PerformanceEvaluation

sketched figure as the query for the

Here

system.This

weintroducethedatabasethatweselecttotestour

phasedescribestheimagessearchingprocess.

system,andwealsocomparethe

Theuserentersaqueryimageforwhichthesyste

systemresultswithsome

mextractscolor,textureandshapefeaturesthefea

CBIRsystems.

turesvectorsofdatabaseimagesarepreviouslyex

Theimagesdatabasethatweusedinourevaluatio

tractedandstored.

nisWANGdatabase.ItisasubsetoftheCoreldata

Similarity

baseof1,000imagesinJPEGformat.

Matching:Usingthesimilaritymetricsdefined

1,000imagedatabasewentthroughourimpleme

forcolor,textureandshape,thesimilaritydistan

ntedsystemtoextractthefeaturesandstoredthe

cesbetweenthequeryimageandthecentroidim

m.TheextractedfeaturesareweightedbyGAan

ageofeachclassarecalculated.

dtheyareusedforclassificationb y

Thesmallestdistance(mostsimilar)willdeterm

HARP

inetowhichtheimagebelongs.Theclasswithth

algorithm.Thelevelofretrievalaccuracyis

esmallestdistanceisreturnedandtheimagesint

fa ct o r

hisclasswillbecomparedwiththequeryimage.

performance.InCBIR,the

Retrieval:The

used

mostmatchingimageswillberetrievedand

performancemeasuresarePrecisionandRecall.

thentheyare

Precisionisdefinedastheratioofthenumberofr

ThefirstN

sorted

in

ascendingorder.

similartarget

to

new

other

inf lu e nc e
most

existing

usingthe

a
t he

commonly

images

etrievedrelevantimagestothetotalnumberofret

withsmallestdistancevaluetothequeryareretri

rievedimages.Thismeansthatprecisionmeasur

evedandshowntotheuser.

estheaccuracyofthe
16

retrieval.Recall

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
isdefinedastheratioofthenumberofretrievedrel

t h e accuracy.

evantimagestothetotal

Here

numberofrelevant

by

usingclusteringpre-

images in thedatabase.The recall measures

processofthedatabaseimagevia

therobustnessoftheretrieval.

algorithmdecreasestheaveragequeryresponset

InCBIR,iftheprecisionscoreis1.0

then

H AR P

ime,

everyimageretrievedbyasearchisRelevant.

thesimilaritysearchtimeforimagematching

Ifthe recallscoreis 1.0 then allrelevant images

and increasestheefficiencyof thesystem.

are

robust.

Comparisonof

u s in g

otherexisting systems

retrievedbythe

search

is

Weevaluatethen e w systemb y

the

new

systemwith

twometricsviz:theRetrievalEffectivenessandth

Foreachclassinthedatabase,werandomlyselect

eRetrievalEfficiency.

ed20imagesasqueries.Sincewehave

a. RetrievalEffectiveness:Aretrievedimageis

3 classesinthedatabase,wehave60queryimages

consideredamatchifitisinthesameclassasth

.Foreachquery,wecalculatetheprecisionandrec

equeryimage.Thesystemworkswellanditre

alloftheretrieval.Theaverageprecisionsandthe

trievesb e t t e r

averagerecallforeachclassbasedonthereturned

resultsovertherandomlyselectedimagesasq

top20imageswererecorded.Moreover thenew

ueries

systemresultiscomparedagainsttheperformanc

by

using

GAand

HARP

a lgo r it h m.

eof three methods.

1.2

weightstoeachfeature

CF

Average Precision

b. RetrievalEfficiency:Byassigningdifferent

to

CTF

0.8

improvetheefficiencywe
GAwitha

HARP

have

a lg o r it hm

used

CSF

0.6

to

CTSF

0.4
elephantClass ID horse

selectoptimumweights offeaturest o g e t

17

dinosaurs

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
Figure 2: Comparison of Precision ofthenew

1.1

Improved K-means

system with various types of features.

The

above

Multi-round RF

figureshowsthatthenew
Average Precision

0.9

systemperformssignificantlybetter in all three

0.8

classes

Interactive GA

except

elephant

clustering-based indexing
technique
New system

0.7

class.Thisresultconfirmsthatafusionofmultipl
efeaturescanincreasetheperformance

0.6

of

0.5

thesystem.
The

below

figure

showsthatthe

0.4

new

Elephant

systemperformssignificantlybetterthanothere

Horse

dinosaurs

Class ID

xisting systemsforallclassesexcept e lephant

Figure3:ComparisonofPrecisionofthen e w
c la ss.
s ystemwiths o m e e xisting Systems
Thisisagoodindicatorfortheeffectivenessofour
1

system.Thereasonbehindthelimitationintwocl

Interactive GA
Improved K-means

assesisthatthoseclasses'imagesareverysimilari

0.8
Average Recall

Multi-round RF

nterm ofthe dominant color, textureandshape

so,our

new

0.6

systemmayconfusedbetween

clustering-based indexing
technique
New system

0.4

them.

0.2
0
Elephant

Horse

dinosaurs

Class ID

Figure4: Comparison ofRecall of the new

18

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
Systemwith some existings ystems

beinggenerated, for finding relatively similar

The

images. In thiswork the GA is combined with

above

figure

showsthatthenew

systemperformssignificantlybetterthanothers

HARP clustering algorithm toimprove the

ystems,forallclasses.Thismeansthatthenew

retrieval accuracy of the system. Getting lower

systemcanretrievemostofdatabaseimagesthat

computational time and retrieving relevant

matchqueryimage.The

new

and accur at e image is possible by using

systemworkswellintheclassificationpartofusin

CBIR. In future we have a proposal to

gGAwith

disseminate the features selections and use

HARP

algorithm.Theaverageprecisiona n d

other distance measures to improve the overall

theaveragerecallincreasedfrom78.1%to88.2%

results.

and

Theefficiencyof

50.4%to69.9%respectively

and

obtainedanaveragereductionin 6.21 seconds.

the

new

system

is

improvedbyconsideringcandidateimages

for

similaritycomputation

not

purpose

i.e.

considering the wholedatabaseimages. A

Conclusion

candidateimageliesinthesameclusterwiththequ

The explosive growth of image data leads to

eryimage the benefit of the clustering process
theneed of research and development of Image
clearly proved the retrievalaccuracy.
Retrieval.CBIR is currently a keen area of
REFERENCES
research in the area of multimediadatabases.
1. V.GudivadaandV.Raghavan,
Various research works had beenundertaken in
the

past

decade

to

design

Content-

efficient
basedImageRetrievalSystems,IEEE

imageretrieval techniques from the image

Computer, vol. 28, no 9, pp18-22,
databases. More prcised retrieval techniques
Sep. 1995.
areneeded to access the large image archives
19

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
2. F.Long,H.Zhang,H.Dagan,andD.

pp-1-11, October 2014,

ISBN: 978-93-84052-11-9.

Feng,FundamentalsofContentBased

5. R.Chang,J.Ho,S.Lin,C.Fannand
ImageRetrieval,MultimediaSignalP
Y.Wang,ANovelContentBasedIma
rocessingBook,Chapter1,SpringergeRetrievalSystemusingKVerlag,BerlinHeidelbergNew York,
meanswithFeatureExtraction,Intern
2003.
ationalConferenceonSystems

and

3. S.Selvam and Dr.S.ThabasuKannan,

Informatics,2012.

Design of an Effective Method for

Image
IJIRAE,

Retrieval, published
International

6. I.El-

Journal of

Naqa,Y.Yang,N.Galatsanos,R.Nishi

Innovative Research in Advanced

kawaandM.Wernick,ASimilarityLe

Engineering, Volume-1, March 2014,

pp.51-56.

arningApproachtoContentBasedImageRetrieval:Applicationto

4. S.Selvam and Dr.S.ThabasuKannan,

An Empirical Review on Image

DigitalMammography,IEEE

Retrieval System by using Relevance

Transactions on Medical Imaging,

Feedback proceeding of International

2009.

Symposium on Research innovation

for quality improvement in Higher

7. B.WANG,X.ZHANG,andN.LI,

Education conducted by Bharathiar

RelevanceFeedbackTechniqueForC

University, Coimbatore, October 2014

and published in Research and Trends

ontent-

in Data mining and Image Processing

BasedImageRetrievalUsingNeuralN

Technologies and Applications,

etworkLearning,ProceedingsoftheF

Bloomsbury publishing India, London,

ifth

New Delhi, New York, Sydney

InternationalConferenceon Machine

2006.

Learning and Cybernetics,Dalian,

8. R.Datta,J.Li,andJ.Wang,Content20

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
BasedImageRetrieval-

ceptsandTechniques,2ndEd.,Morgan

ApproachesandTrends of theNew

KaufmannPublisher, 2006.

Age,ACMComputing Surveys,

10. P.JeyanthiandV.JawaharSenthil
Kumar,ImageClassificationbyK-

vol. 40, no. 2, pp. 1-60,April 2008.

meansClustering.Advances

9. J.HanandM.Kambr,DataMiningCon

inComputational
Sciences and Technology, 2010.

1. FIRST AUTHOR PERSONAL PROFILE:

Prof. S.SELVAM M.Sc., M.Phil, has been working as Assistant Professor and Head,
Department of Computer Application, N.M.S.S.Vellaichamy Nadar College, Nagamalai, Madurai19, Tamilnadu, India. He has more than 13 years of teaching experience. He has published three
research papers in various refereed International/National Level Journals/Proceedings and
Conference. His research paper also published in the book titled Research and Trends in Data
mining and Image Processing Technologies and Applications, Bloomsbury publishing India PVT,
LD, London, New Delhi, New York, Sydney pp-1-11, October 2014, ISBN: 978-93-84052-11-9.
Under his guidance two M.Phil scholar were awarded. His area of interest is Digital Image
Processing.

21

International Journal of Emerging Technology in Computer Science & Electronics (IJETCSE)

ISSN: 0976-1353 Volume 12 Issue 2 JANUARY 2015.
2. SECOND AUTHOUR PERSONAL PROFILE:

Prof.Dr.S.ThabasuKannan has been working as Professor and Principal in Pannai

College of Engineering and Technology, Sivagangai and rendered his valuable services for more
than two decades in various executive positions. He has published more than 50 research level
papers in various refereed International/National level journals/proceedings. He has authored 11
text/reference books on the information technology. He has received 11 awards in appreciation
of his excellence in the field of research/education. He has visited 5 countries to present his
research papers/articles in various foreign universities. He has been acting as consultant for
training activities at Meenakshi Trust, Madurai. His area of interest is Big data applications for
bioinformatics applications. Under his guidance 8 Ph.d scholars pursuing and more than 150
M.Phil scholars were awarded. His several research papers have been cited in various citations.

22