Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Save to My Library
Look up keyword
Like this
4Activity
0 of .
Results for:
No results containing your search query
P. 1
Off-Line Handwritten Signature Retrieval using Curvelet Transforms

Off-Line Handwritten Signature Retrieval using Curvelet Transforms

Ratings: (0)|Views: 263 |Likes:
Published by ijcsis
In this paper, a new method for offline handwritten signature retrieval is based on curvelet transform is proposed. Many applications in image processing require similarity retrieval of an image from a large collection of images. In such cases, image indexing becomes important for efficient organization and retrieval of images. These papers address this issue in the context of a database of handwritten signature images and describes a system for similarity retrieval. The proposed system uses a curvelet based texture features extraction .The performance of the system has been tested with an image database of 180 signatures. The results obtained indicate that the proposed system is able to identify signatures with great with accuracy even when a part of a signature is missing.
In this paper, a new method for offline handwritten signature retrieval is based on curvelet transform is proposed. Many applications in image processing require similarity retrieval of an image from a large collection of images. In such cases, image indexing becomes important for efficient organization and retrieval of images. These papers address this issue in the context of a database of handwritten signature images and describes a system for similarity retrieval. The proposed system uses a curvelet based texture features extraction .The performance of the system has been tested with an image database of 180 signatures. The results obtained indicate that the proposed system is able to identify signatures with great with accuracy even when a part of a signature is missing.

More info:

Published by: ijcsis on Dec 04, 2010
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

12/09/2013

pdf

text

original

 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
8
 , No.
8
 , 2010
 
Off-Line Handwritten Signature Retrieval using Curvelet Transforms
M. S. Shirdhonkar Dept. of Computer Science and Engineering,B.L.D.E.A’s College of Engineering and TechnologyBijapur, IndiaE-mail:ms_shirdhonkar@rediffmail.comManesh KokareDept. of Electronics and Telecommunication,S.G.G.S Institute of Engineering and TechnologyNanded, IndiaE-mail: mbkokare@sggs.ac.in
 Abstract 
 —— In this paper, a new method for offline handwrittensignature retrieval is based on curvelet transform is proposed.Many applications in image processing require similarityretrieval of an image from a large collection of images. In suchcases, image indexing becomes important for efficientorganization and retrieval of images. These papers address thisissue in the context of a database of handwritten signature imagesand describes a system for similarity retrieval. The proposedsystem uses a curvelet based texture features extraction .Theperformance of the system has been tested with an imagedatabase of 180 signatures. The results obtained indicate that theproposed system is able to identify signatures with great withaccuracy even when a part of a signature is missing.
 Keywords- Handwritten recognition, Image indexing, Similarityretrieval, Signature verification, Signature identification.
I. I
 NTRODUCTION
 
(H 
 EADING
1) A.Motivation
A signature appears on many types of documents such as bank cheques in daily life and credit slips, thus signature has agreat importance in a person’s life. Automatic bank cheque processing is an active topic in the field of document analysisand processing. Signature validity confirmation of differentdocument is one of the important problems in automaticdocument processing. Now a days, person identification andverification are very important in security and resource accesscontrol. For this purpose the first and simple way is to usePersonal Identification Number (PIN), but PIN code may beforgotten. Now an interesting method to identification andverification is biometric approach [1]. Biometric is a measurefor identification that is unique for each person. Always biometric is together with person and cannot be forgotten. Inaddition biometric usually cannot be misused.Handwritten signature retrieval is still a challenging work in the situations of a large database. Unlike fingerprint palm print and iris, signatures have significant amount of intra classvariations making the research even more compelling. Thisapproach with the potential applications of signaturerecognition/verification system optimized with efficientsignature retrieval mechanism.
 B.Related works.
Signature verification contain two areas: off-line signatureverification ,where signature samples are scanned into imagerepresentation and on-line signature verification, wheresignature samples are collected from a digitizing tablet whichis capable of pen movements during the writing .In our work,we survey the offline signature identification and retrieval . In2009, Ghandali and Moghaddam have proposed an off-linePersians signature identification and verification based onImage registration, DWT (Discrete Wavelet Transform) andfusion. They used DWT for features extraction and Euclideandistance for comparing features. It is language dependentmethod [1]. In 2008, Larkins and Mayo have introduced a person dependent off-line signature verification method that is based on Adaptive Feature Threshold (AFT) [2]. AFTenhances the method of converting a simple feature of signature to binary feature vector to improve its representativesimilarity with training signatures. They have usedcombination of spatial pyramid and equimass sampling gridsto improve representation of a signature based on gradientdirection. In classification phase, they used DWT and graphmatching methods. In another work, Ramachandra et al [3],have proposed cross-validation for graph matching based off-line signature verification (CSMOSV) algorithm in whichgraph matching compares signatures and the Euclideandistance measures the dissimilarity between signatures.In 2007, Kovari et. al presented an approach for off-linesignature verification, which was able to preserve and takeusage of semantic information[4].They used position anddirection of endpoints in features extraction phase. Porwik [5]introduced a three stages method for offline signaturerecognition. In this approach the hough transform ,center of gravity and horizontal-vertical signature histogram have beenemployed, using both static and dynamic features that were processed by DWT has been addressed in[6].The verification phase of this method is based on fuzzy net using the enhancedversion of the MDF(Modified Direction feature)extractor has been presented by Armand et.al [7].The different neuralclassifier such as Resilient Back Propagation(RBP), Neuralnetwork and Radial Basis Function(RBF) network have beenused in verification phase of this method. In 1995, Han andSethi [8], described offline signature retrieval and use a set of geometrical and topological features to map a signature onto
48http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
8
 , No.
8
 , 2010
2D strings. We have proposed an offline signature retrievalmodel based on global features.The main contribution of this paper is that, we have proposedoff-line handwritten signature retrieval using curvelettransform, In retrieval phases Canberra distance measure isused. The experimental results of proposed method weresatisfactory and found that it had better results compare withrelated works. The rest of paper is organized as follows: Insection II, discusses the feature extraction phase. The signatureretrieval is presented in section III. In section IV, theexperimental results and finally section V concludes thework.II.FEATURE EXTRACTION PHASEThe major task of feature extraction is to reduce image data tomuch smaller amount of data which represents the importantcharacteristic of the image. In signature retrieval, edgeinformation is very important in characterizing signature properties. Therefore we proposed to use the curvelettransform. The performance of the system is compared withstandard discrete wavelet transform which capturesinformation in only three directions.
 A.Discrete Wavelet Transform
The multi resolution wavelet transform decomposes asignal into low pass and high pass information. The low passinformation represents a smoothed version and the main bodyof the original data. The high pass information represents dataof sharper variations and details. Discrete Wavelet Transformdecomposes the image into four sub-images when one level of decomposing is used. One of these sub-images is a smoothedversion of the original image corresponding to the low passinformation and the other three ones are high pass informationthat represents the horizontal, vertical and diagonal edges of theimage respectively. When two images are similar, their difference would be existed in high-frequency information. ADWT with N decomposition levels has 3N+1 frequency bandswith 3N high-frequency bands [9], [10]. The impulse responsesassociated with 2-D discrete wavelet transform are illustrated inFig. 1 as gray-scale image. 
Fig. 1.Impulse response of 0
0
, 90
0
and
±
45
0
of DWT
 B.Curvelet Transform
Recently, Canes and Donoho developed a newmultiscale transform which they called the curvelet transform.Motivated by the needs of image analysis, it was neverthelessfirst proposed in the context of objects f(x1, x2) defined on thecontinuum plane (x1, x2) € R 
2
.The transform was designed to represent edges and other singularities along curves much more efficiently thantraditional transforms, i.e. using many fewer coefficients for agiven accuracy of reconstruction. Roughly speaking, torepresent an edge to squared error 1/N requires 1/N wavelets.The curvelet transform, like the wavelet transform, is amultiscale transform, with frame elements indexed by scale andlocation parameters. Unlike the wavelet transform, it hasdirectional parameters, and the curvelet pyramid containselements with a very high degree of directional specificity. Inaddition, the curvelet transform is based on a certainanisotropic scaling principle which is quite different from theisotropic scaling of wavelets. The elements obey a specialscaling law, where the length of the support of a frameelements and the width of the support are linked by the relationwidth ≈ length
2
.see details in [11].
C.Feature Database Creation
To construct the feature vectors of each handwrittensignature in the database using DWT and curvelet transformrespectively. The Energy and Standard Deviation (STD) werecomputed separately on each sub band and the feature vector was formed using these two parameter values. The Energy
 E 
and Standard Deviation
σ   
of
th
sub band iscomputed as follows
( )
= =
×=
i N  j
ji N  E 
1 1
,1
(1)
( )( )
2121 1
,1
×=
= =
 N i  j
ji N 
 µ σ  
(2)Where
),(
ji
is the
th
wavelet-decomposedsub band,
Mx
is the size of wavelet decomposed sub band, and
 µ 
is the mean of the
th
sub band. Theresulting feature vector using energy and standard deviation are
[ ]
n E 
 E  E   f  
...
21
=
and
[ ]
n
 f  
σ  σ  σ  
σ  
...
21
=
respectively. So combinedfeature vector is
[ ]
nn
 E  E   f  
......
2121
σ  σ  σ  
σµ 
=
(3)III. OFFLINE H
ANDWRITTEN
SIGNATURE R 
ETRIEVAL
 PHASEThere are several ways to work out the distance betweentwo points in multidimensional space. The most commonlyused is the Canberra distance measure. It can be considered theshortest distance between two points. We have used Canberradistance metric as similarity measure. If x and y are the featurevectors of the database and query signature, respectively, x andy have dimension d, then the Canberra distance is given byCanb (x, y)
+=
=
iiiii
 y x y x
1
 (4)
49http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol.
8
 , No.
8
 , 2010
Algorithm 1: Offline Handwritten Signature RetrievalInput: Test signature: StFeature database: FVOutput: Distance vector: DistHandwritten signature retrievalBeginCalculate feature vector of test signature usingDWT and curvelet transformFor each fv in FV doDist= Calculate distance between test signatureand fv using (4)sort DistEnd for Display the top signature from dist vector.EndIV.EXPERIMENTAL RESULTS
 A.Image Database
The signatures were collected using either black or blue ink (No pen brands were taken into consideration), on a white A4sheet of paper, with eight signature per page. A scannesubsequently digitized the eight signatures, contained on each page, with a resolution in 256 grey levels. Afterwards theimages were cut and pasted in rectangular areas of size256x256 pixels. Sample signature database for 16 persons areshown in Fig.2. A group of 16 persons are selected for 12specimen signatures which make the total of 16x12=192signature database. Fig.2. Sample Signature Images Database
Query imageFig.2. Sample Handwritten Signature Images Database
 B.Retrieval Performance
For each experiment, one image was selected at random as thequery image from each writer and thus retrieved images wereobtained. For performance evaluation of the signature imageretrieval system, it is significant to define a suitable metric.Two metrics are employed in our experiments as follows.
 Recall 
=
 signaturesrelevant of   Number  retrieved  signaturesrelevant of   Number 
(5)
 Precision
=
retrieved  signaturesof   Number  retrieved  signaturesrelevant of   Number 
(6)Results correspond to precision and recall rate for a Top1, Top2, Top 5, Top 8, Top 10, and Top 12. The comparativeretrieval performance of the proposed system is shown inTable 1. 
Table1: Average Retrieval Performance
Discrete waveletTransformCurvelet Transform
Number of Top matches
Precision%Recall%Precision%Recall%
Top 1
10081008
Top 2
8012.696.615.4
Top 5
66.728.99236.7
Top 8
55.837.273.348.5
Top 10
51.343.470.759.0
Top 12
47.847.566.0465.2
Retrieval performance of the proposed method is comparedusing DWT transform technique. We evaluated the performance in terms of average rate of retrieving images asfunction of the number of top retrieved images. Fig.3 showsgraph illustrating this comparison between DWT and curvelettransform according to the number of top matches consideredfor database. From Fig. 3, it is clear that the new method issuperior to DWT. To retrieve images from the database thosehave a similar writing style to the original request. In Fig. 4,retrieval example results are presented in a list of imageshaving a query image.
50http://sites.google.com/site/ijcsis/ISSN 1947-5500

Activity (4)

You've already reviewed this. Edit your review.
1 thousand reads
1 thousand reads
1 hundred reads

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->