Signature Verification using DCT

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056
Volume: 04 Issue: 06 | June -2017 www.irjet.net p-ISSN: 2395-0072
© 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1437
SIGNATURE VERIFICATION USING DCT
Prathiba M K1, Sushmitha H L2, Chandan B3, Karthik B K4 and Manohar H P 5
1Professor , Dept. of Electronics and Communication Engineering, ATMECE Mysore, Karnataka, India
2,3,4,5UG Students Dept. of Electronics and Communication Engineering, ATMECE Mysore, Karnataka, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Signatures have been considered a typical form
of authentication in our society for hundreds of years.
Signature is a simple, concrete expression of the unique
variations in human hand geometry. The way a person signs
his or her name is known to be characteristic of that
individual. A signature verification system must be able to
detect forgeries, and, at the same time, reduce rejection of
genuine signatures. Signatures are subject to intra-personal
variations. Hence, a signature verification system is feasible
only if the system is insensitive to intra-personal variability,
but sensitive to interpersonal variability. Even when
insensitive to intra-personal variations, the system must
possess the discriminating power to foil skillful forgers.
Key Words: (Size 10 & Bold) signature verification,
interpersonal variability, intra-personal variability,
Discrete Cosine Transform, Feature extraction etc
(Minimum 5 to 8 key words)…
1. INTRODUCTION
Human signature is a biometric measure of person's
identification. In many sectorslikebanks,official documents,
receipts etc. handwrittensignatureareverifiedtosecure and
identify the concerned person. Each individual has his own
signature different with others but in most of the time it is
not possible to produce the replica of the signature. The
signature verification problem aims to minimize the
intrapersonal differences. Signature verification can be
categorized into following two parts: online and offline
.Online handwritten signature verification deals with
automatic conversion of characters which are written on a
special digitizer, tablet PC or PDA, Whereina sensorpicksup
the pen -tip movements as well as pen -up/pen-down
switching.
An on-line signature verification system based on local
information and a one-class classifier, i.e. the Linear
Programming Descriptor classifier (LPD), was presented by
Nanni and Lumini [1]. The authors investigated and
described how the information was extracted as time
functions of various dynamic properties of the signatures,
and then the discrete 1-D Wavelet Transform (WT) was
performed on these features. The DCT was used to reduce
the approximation coefficients vector obtained by WT to a
feature vector of a given dimension. Fabregas and Faundez-
Zanuy [2]: have presented a new system for on-line
signature verification based on DCT feature extraction with
discriminability feature selection. They performed a
complete set of simulations with the largest available online
signature database, MCYT, whichconsistsof330peoplewith
genuine and skilled forgeries performed by five other
different users.
Dimauro et.al.[3] in their study of local features stability
have introduced a warping functionthatallowsmtonpoints
to be matched. However, such an approach may not be
practical in cases of extreme values of features.
Lei and Govindaraju[4] analyzed the consistency and
discriminative power of on-line features using a distance-
based measure that is optimized for each feature of study.
They view signature verification as a one category
classification problem, and use the distances between
features to distinguish them rather than the feature values
themselves.
1.1 Motivation
Online handwrittensignatureverificationsystemisbased
on the extracting the featuresof signaturefromthesignature
pad and consider as interpersonal features that are not
statistically stable. These are used for comparison and
verifications between different signaturesandenhancingthe
interpersonal variations and classifier is used to verify
signatureand provide resultasgenuineorforgeryandalsoto
present a simple and effective approach for an online
signature verification system. Finally, complete content and
organizational editing before formatting. Please take note of
the following items when proofreading spelling and
grammar:
1.2 Discrete Cosine Transform (DCT)
Online signature verification has proved to be more
accurate than the offline signature verification. The DCT has
the ability to compactly represent an online signature
verification using a fixed number of coefficients which leads
to fast matching algorithm.
A DCT coefficient is the compression of feature data,
which reduces the elapsed time and storage space for
training and verification processes. The time for DTW and
HMM methods is very long. From these advantagesofDCT,it
motivated us to do online signature verification based on
DCT.
2. SIGNATURE VERIFICATION SYSTEM
The proposed online handwritten signature verification
system consists the following blocks as show in Fig. 1

Fig -1: Block diagram of signature verification system
During signature acquisition phase signatures are acquired
by using Wacom tablet the 125 signatures are taken from 25
signers at least 5 samples each user. The features are
extracted from these signatures and stored in the database.
Testing phase is carried out same as training phase but in
testing phase we do feature extraction is only for specified
sample of specified user. If, the tested signature features
matches with the features of the training phase of the
respective user then the tested signature will treated as
genuine signature.
During the training and verification phase, test signature to
be compared against the claimed user’s reference set
signature. In order to match two signatures, the following
stages are applied:
A basis feature matrix is calculated from the reference
signatures of each signer.Forthis,wecalculatetheminimum,
mean, maximum and sum values for each DCT Coefficient in
the reference set. The distance between the featurematrixof
reference and test signatures computes. The matrix of the
featuredistance isalso normalizedtomaptoavaluebetween
0 and1. Genetic algorithm is used for verification process.
3. RESULTS
The signature is acquired using tablet is as shown in Fig.2
Fig -2: User signature
The user signature is varying with respect to horizontal
direction i.e. X-coordinate and the DCT of X is shown in Fig.
3.
Fig -3: Plotting of theta and its DCT of signature
The user signature is varying with respect to vertical
direction i.e. Y-coordinate and the DCT of Y is shown in Fig.
4.
The value of theta represents the movement of pen with
respect to X-axis and its corresponding DCT is plotted as
shown in Fig.5.

Genetic algorithm is used for classification of signatures,
which makes the verification performancemoreaccurate. To
evaluate the experiments, False Acceptance Rate (FAR) and
False Rejection Rate (FRR) are determined. FAR represents
the probability that a false match occurs, while FRR
represents the probability that a false rejection occurs.
Genuine and forged signatures of each user are fed to
database for comparison which provides two possible
results and are as follows.
When genuine signatureiscomparedwithgenuinesignature
of users and the result obtained is “Genuine signature” as
shown in Fig. 6.
Fig -6: output of genuine signature
Genuine signature is compared with forged signature of
users and the result obtained is “forged signature” asshown
in Fig.7.
Fig -7: output of genuine signature
3. CONCULSION
Although signature verification is not one of the safest
biometric solutions, the use of it in business practices is still
justified. Moreover, signature verification has a very
promising future. One major drawback is that humans are
not consistent when signing their signatures. In this paper, a
robust signature verification system has been proposed,
based on DCT coefficients and the Genetic Algorithm. Our
proposed method can extract basic dynamic features from
signature time signals, and compress signature data, while
keeping the rough form and basic information of signatures.
Especially in the context of skilled forgery, where inter-
personal variability in the number of features becomes
negligible, an effective analysis of features based on time
signals is essential for attainment of a suitable performance.
Ideally we need features that are stable, i.e. do not change
very much between different genuine signatures, and which
are hard to forge. For attaining this purpose, capturing
signals via a tablet digitizer and using the extracted
dynamics information has been considered, mainly in the
form of simple parameters of DCT coefficients. These
features are efficient and experimental results confirm that
the proposed method is promising. The summation
condition in the verification process guides the system
towards an accurate decision.
REFERENCES
[1] Nanni, L. and Lumini, A., “A novel local on-line signature
verification system,” Pattern Recognition Letters, 2008.,pp.
559–568.
[2] Fabregas, J. and Faundez-Zanuy,., “On-line signature
verification system with failure to enroll management,”
Pattern Recognition Letters, 2009., pp. 2117–2126.
[3] Dimauro, G., Modugno, R., Pirlo, G. and Sarcinella, L.,
“Analysis of stability in hand-written dynamic signatures,”
Proceedings of the 8th International Workshop onFrontiers
in Handwriting Recognition, IWFHR,2002, pp. 259–263
[4] H. Lei, V. Govindaraju “A comparative study on the
consistency of features in on-line signature verification”
Pattern Recognition Letters, 26 (15) (2005),pp.2483– 2489

Signature Verification using DCT

More Related Content

What's hot (7)

Similar to Signature Verification using DCT (20)

More from IRJET Journal (20)

Recently uploaded (20)

Signature Verification using DCT