SlideShare a Scribd company logo
1 of 6
Download to read offline
Dynamic Signature Forgery and
         Signature Strength Perception Assessment

 Stephen Elliott & Adam Hunt
 Purdue University



 ABSTRACT                                                                                              include the fact that a signature is learned over time (and
                                                                                                       evolves over time as the owner and his or her handwriting
    Dynamic signature verification has many challenges                                                 matures), it contains variant measures (such as pressure,
associated with the creation of the impostor dataset. The                                              speed, etc., that can be changed), can be changed by the
literature discusses several ways of determining the                                                   owner (depending on the ceremony of the transaction), and
impostor signature provider, but this takes a different                                                may have several versions (for example, at work and home
approach - that of the opportunistic forger and his or her                                             may have different signatures).
relationship to the genuine signature holder. This                                                         A discussion of DSV invariably raises a number of
examines the accuracy with which an opportunistic forger                                               concerns. The first concern is that people acknowledge their
assesses the various traits of the genuine signature, and                                              failure to sign consistently, and the second is that most
whether the genuine signature holder believes that his or                                              people have attempted, irrespective of degree of success, to
her signature is easy to forge.                                                                       forge someone's signature at some time. In fact, a straw poll
                                                                                                      conducted in a class of 80 undergraduate college students
INRODUCTION                                                                                           revealed that at least 90% of them have attempted to forge a
                                                                                                      signature at one time. When asked for more details about the
    Dynamic signature verification (DSV) has long been used                                           forgery attempt, in the majority of cases, the subject of the
to authenticate individuals based on their signing                                                    attack is someone who is known to the forger (typically a
characteristics, such as speed, pressure, and graphical output.                                       parent or close relative) and the signature is easily available.
Approaches to DSV have been discussed in detail in the                                                In these cases, it is more than likely that the forger has had
literature. Popular applications, such as document                                                    several chances to practice the signature and that the
authentication, financial transactions, and paper-based                                               signature is not rigorously checked by the receiver of the
transactions have all, at one time, used the signature to                                             document being forged. These two conditions correlate to a
convey the intent to complete a transaction [1, 2]. DSV is a                                          low chance of getting caught - this is the scenario of the
subset of a larger science called biometrics. Biometrics aims                                         opportunistic forger.
to authenticate an individual based on either behavioral or                                               Another important consideration has to do with whether a
physiological traits, (or a combination of both), including                                           genuine signature holder believes that his or her signature is
face recognition, iris recognition, and fingerprint recognition,                                      difficult to forge, and whether the imposter also believes that
to name a few. Many of these modalities are made up of both                                           to be the case. The approach, proposed herein, is to
behavioral and physiological attributes, with various                                                 understand whether the impostor can actually make
proportions of each. Within the continuum, the signature is a                                         well-informed decisions on the measurable variables of the
strong behavioral biometric. The signature's unique traits                                            genuine signature. For example: Can the forger determine the
make it harder to test and evaluate than some of the other                                            speed of the signature, as well as the handedness of the
behavioral biometrics, such as voice or face recognition.                                             genuine signer? If the forger can determine these most basic
Challenges to testing and definitively evaluating the signature                                       of attributes, then he or she might then achieve some level of
                                                                                                      success to forge some of the additional variables within DSV.

                                                                                                      VARIABLE CHARACTERISTICS OF THE DSV
Author's Current Address:
S. Elliott, Ph.D., and A. Hunt, Associate Professor, Department of Industrial Technology,
College of Technology, Purdue University, 401 N. Grant Street, West Lafayette. IN 47906,
                                                                                                         DSV's numerous variables are calculated using the input
USA.                                                                                                  gathered from a digitizer. These variables include x and y
Based on a presentation at Carnahsan 2006.                                                            (Cartesian) coordinates, pressure (p) or force, and time (t)
0885/8985/08/ USA $25.00 0 2008 IEEE                                                                  [3]. This output from the digitizer is used to create the global


IEEE A&E SYSTEMS MAGAZINE, JUNE 2008                                                                                                                                 1
                                                                                                                                                                     13


            Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
Table 1. DSV Studies & Error Rates

                                                                                                                                 Error Rates
      Study Authors                                                                         Type I (FAR)                           Type 11             (FRR)EER


     Achemlal, Mourier, Lorette & Bonnefoy (1986)                                                 11.0%                               8.0%
     Beatson (1986)                                                                                1.0%                               2.0%
     Bechet (1984)                                                                                 5.0%                               5.0%
     Bonnefoy & Lorette (1981)                                                                     0.6%
     Bault & Plamondon (1981)                                                                      1.2%                               1.0%
     Collantier (1984)                                                                            3.5%
     Crane & Ostrem (1983)                                                                         1.5%                              1.5%
     Debruyne (1985)                                                                              3.0%                               2.0%
     Hale & Pagnini (1980)                                                                         1.5%                            1.2/2.5%
     Herbst & Liu (1979)                                                                           1.7%                              0.4%
     Herbst & Liu (1979)                                                                          2.4%                               3.4%
     Ibrahim & Levrat (1979)                                                                      19.0%                              5.5%
     Lam & Kamis (1979)                                                                           0.0%                               2.5%
     Lorette (1983)                                                                               6.0%
     Mital, Hmn & Leng (1989)                                                                     0.0%                                0.0%
     Parizeau & Ilamondon (1989)                                                                  4.0%
     Sato & Kogure (1982)                                                                         1.8%                                0.0%
     Worthington, Chainer, Williford
        & Gundersen (1985)                                                                      1.8%                             0.28-2.33%
     Zimmerman & Varady (1985)                                                                30-50%                               4-12%
     Cordelia, Foggia, Sanson & Vento                                                       0.03-20.82%                             5.7%
     Wirtz (1997)                                                                                                                                         10%
     Dimauro, Impedovo & Pirlo (1993)                                                             1.7%                                1.2%
     Naiwa (1997)                                                                                                                                          3%
     Naiwa (1997)                                                                                                                                          2%
     Nalwa (1997)                                                                                                                                          5%
     Mingming & Wijesoma. (2000)                                                                                                                           5%
     Mingming & Wijesomna (2000)                                                                                                                           7%
     Mingining & Wijesoma (2000)                                                                                                                           9%
     Hamilton, Whelan, McLaren & Macintyre (1995)                                              34.0%                                26.0%
     Hamilton, Whelan, McLaren & Macintyre (1995)                                              22.0%                                18.0%
     Hamilton, Whelan, McLaren & Maclntyre (1995)                                              12.0%                                10.0%
     Hamilton, Whelan, McLaren & Macintyre (1995)                                               7.0%                                 6.0%
     Martens & Clausen (1997)                                                                   1.5%                                 1.3%
     Chang, Wang & Suen (1993)                                                                  2.0%                                 2.5%
     Higashino (1992)                                                                           8.0%                                 0.6%
     Minot & Gentric (1992)                                                                     2.0%                                 4.0%
     Lucas & Damper (1990)                                                                      5.6%                                 4.5%
     Tseng & Huang (2002)                                                                    12.5-28.8%                           5.0-12.5%
     Lee, Berger & Aviczer (1996)                                                               1.0%                                20.0%
     Lee, Berger & Aviczer (1996)                                                               5.0%                                20.0%
     Lee, Mohankrshnan & Paulik (1998)                                                          0.9%                                 0.7%
     Han, Chang, Hsu & Jeng (1999)                                                              4.0%                                 7.2%
     Komiya & Matsumoto (1999)                                                                                                                             2%
     Cardot, Revenu, Victorri & Revillet (1993)                                                   2.0%                               4.0%
     Cardot, Revenu, Victorri & Revillet (1993)                                                   0.9%                               7.4%



14                                                                                                                                  IEEE A&E SYSTEMS MAGAZINE, JUNE 2008
                                                                                                                                   14


     Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
and local features described in various accounts [4, 51. These                                    created by distorting real signatures through the addition of
global and local features, derived from the basic set of data                                     low-level noise and dilation or erosion of the various
that a digitizer provides, vary significantly across algorithms.                                  structures of the signature. [14] motivated the forgers by
[6] outline 61 features, [7] note over 40 features, [8] discuss                                   offering a cash reward. [15] examined people's signatures
44 features. In [2], describes the x and y coordinates of the                                     over a four-month period to assess variability over time. In
pen's motion. Also [9], observes that the temporal                                                the Signature Verification Competition, genuine signers
"characteristics the production of an on-line signatureare
                  of                                                                              created signatures other than their own [16]. In [17], the
the key to the signature'sverification" (p. 5). From these                                        author uses a number of different methodologies to generate
many approaches and features, the number of variables                                             the impostor distribution, with the majority of impostors
associated with a dynamic signature can be synthesized to                                         using some form of practice. In [ 18], the authors defined
several major statistical features. These major statistical                                       three different levels of forgeries: the simple, statically
features include pressure, time, horizontal, and vertical                                         skilled, and timed (p. 643). [19] used signatures that "on
components of position, velocity, acceleration, and force, all                                    casual visual inspection would pass as authentic" (p. 201).
measured against time. An alternative approach to                                                 [20] provides three characteristics of forgery: the random
characterizing a signature involves analysis of the "stroke,"                                     forgery, defined as one that belongs to a different writer of
that is, the up and down movements of the pen on the                                              the signature model; simple forgery, represented by a similar
digitizer. The many dynamic traits collected by the digitizer                                     shape consistency with the genuine signer's shape; and the
during the act of signing are said to make an impostor                                            skilled forgery (p. 2).
signature easier to detect than that of a traditional
paper-based impostor signature.
                                                                                                  PERCEIVED STRENGTH OF SIGNATURE (PSS)
HIMPOSTOR SIGNATURE GENERATION
                                                                                                     The purpose of this paper is to assess the basic attack on a
                                                                                                  signature by an opportunistic forger and to determine the
    Impostor datasets are created in numerous ways and have
                                                                                                  perceived strength of the signature (PSS). PSS is a concept
the effect of changing the respective performances of
                                                                                                  that indicates that an opportunistic forger will not forge a
algorithms. This change can be done through the different                                         signature that is difficult to forge, as their success at the
generation of the impostor signatures. A review of the                                            point-of-sale may be not as high as the forgery of an easy
literature shows various performance results from several                                         signature. This is more the trademark of an opportunistic
studies, all of which have different methodologies for
                                                                                                  forgery than of a more sophisticated attack on the signature,
collecting impostor signature datasets. Table 1 outlines the
                                                                                                  as outlined in previous research. For this study, an
various studies and their respective error rates (false accept,
                                                                                                  opportunistic forger is analogous to an opportunistic thief,
false reject, and the equal error rate where appropriate).                                        that is, one who works on his or her own without any
    The variances in error rates shown in Table 1 (0% to 50%                                      equipment [211]. This definition is further enhanced by the
false accept rate and 0% to 20% false reject rate) can be                                         absence of occasion to practice forging the signature. The
explained by a number of factors, one of which has to do
                                                                                                  study outlines a basic truth involving the genuine signature
with how an impostor signature dataset is created. [10]'s
                                                                                                  owner's perception of the strength of their signature and tries
study is particularly interesting. This study had a database
                                                                                                  to understand whether the owner of a genuine signature has
consisting of 293 genuine signatures and 540 forgery
                                                                                                  the same or different perception of the signature than that of
signatures from eight individuals. Although the study did not                                     the forger.
explain how the forgery took place (in terms of training,                                            In order to understand the basic truth of the perceived
payment, etc.), eight individuals created the impostor dataset.                                   strength of the signature, each of the genuine signature
[11] study dataset was comprised of 496 original signatures                                       owners was asked for the following information about their
from 27 people. Each person signed I11 to 20 times. The
                                                                                                  signature:
database contained 48 forgeries that "fulfill the requirement
on the visual agreement and the dynamic similarity with the
original signature" (p.5). [12] trained the algorithm using                                                1. How easy their signature was to forge (rated on
250 signatures per writer; of these 250 signatures, 100 were                                                  a Likert scale).
authentic signatures and 150 were random forgeries,
classified as the genuine signatures of other writers. [ 13] used                                         2. How fast or slow they signed their signature
27 people in their study, with the participants writing their                                                (rated on a Likert scale from slow to fast).
own signature. The study also used 4 people who imitated the
signatures of these 27 people. Unfortunately, no further                                                  3. Handedness (right- or left-handed;
information is provided on the selection of the impostor or on                                               ambidexterity was not an option captured by the
what knowledge the forgers possessed in order to forge the                                                   survey).
signatures.
   [1] used genuine signatures from other individuals as                                            The objective in obtaining these three pieces of
forgeries. In addition, a group of synthesized signatures was                                    information was to assess whether the forger was able to


IEEE A&E SYSTEMS MAGAZINE. JUNE 2008                                                                                                                            15

         Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
predict the speed and forgeability of genuine signatures,                                                    2. How fast or slow was the genuine signature
which was typically centered on the dynamic traits of speed                                                     estimated to be made (rated on a Likert scale
or velocity, time, and graphical outline or complexity of the                                                   from slow to fast).
shape. Furthermore, subjects were asked to sign their name
on a digitizer so that feature variables could be extracted to                                               3. What was the handedness of the genuine subject
estimate whether there were any correlations between                                                            (right- or left-handed; ambidexterity was not an
variables and the respective PSS categories.                                                                    option captured by the survey).

              Table 2. T-Test for Difficulty Groups                                                    The results were analyzed statistically to determine
                                                                                                    whether any significant differences existed between the
                                                                                                    genuine signature owners and the impostors regarding their
                                                                                                    assessments of the signature.

                                                                                                    RESULTS

                                                                                                        For the genuine dataset, a total of 60 subjects participated,
                                                                                                    of which 1 was female and 59 were males. Of these 60
                                                                                                    subjects, 36 signed the digitizer. The remainder did not sign
                                                                                                    (or dropped out of the study). This represents a retention rate
                                                                                                    in the study of 60%. For the impostor dataset, there were 9
                                                                                                    individuals who ranked the genuine signatures using the
                                                                                                    parameters previously described.
                                                                                                       A t-test was used to determine whether the mean of the
METHODOLOGY                                                                                         genuine and impostor groups were statistically significant
                                                                                                    from each other with regards the three questions posed: the
     In order to assess PSS, two separate groups were                                               rank of the perceived level of difficulty, velocity, and
 organized. The dataset of genuine signatures was collected                                         handedness. A level of 0.05 was selected for determining
 from consent forms signed by the genuine signature owners.                                         statistical significance. In the study, the data were normally
 The consent forms were used to maintain a level of                                                 distributed, and there were no outliers. The data were ranked
 ceremony, since the consent form is a document that requires                                       from 1 to 5, with 1 being easy to forge and 5 being difficult
 a signature with a level of intent, as opposed to a random                                         to forge.
 signature with no intent. This signature was subsequently
 used as the target signature. In order to estimate whether any                                                               Table 3. Signature Speeds
 of the dynamic signature verification variables were the same
 for each group (those who ranked their signature within the
 same Likert classification), each subject signed his or her                                                         Speed                               p-value
 name on a digitizer three times. In order to obtain a
 consistently precise signature, the study utilized an Interlink
 Electronics ePad-ink ProTM1 device, which has 100-400 reports                                                           1                                 0.002
 per second and 300 dots per inch [22].
    The device was connected to forensic signature software                                                             2                                  0.011
to extract the raw data from the digitizer, but the subjects                                                            3                                  0.060
could not see the signature or the information on the PC
monitor as they signed. The digitizer provided an inked                                                         No group 4                                  N/A
display of the signature as the subject signed his or her name.                                                         5                                  0.035
The three signatures were then processed and the resultant
variables averaged across the signatures.
    The impostor group consisted of individuals other than
                                                                                                      When assessing the PSS, a t-test result shows parallels
those who owned the original signatures. Members of the
impostor group were asked for information about what they                                          between genuine subjects and forgers when the level of
                                                                                                   difficulty was assessed as "neutral." Likewise, the difference
observed while looking at the signed consent form of each
                                                                                                   in the "very difficult" ranking has a p-value of 0.053. Other
individual in the genuine group:
                                                                                                   categories (levels 1, 2, and 4 in the Likert scale) exhibited
                                                                                                   significantly different means. It is difficult to determine
      1. How easy the genuine signature was to forge                                               whether the groups were statistically significant. Further
         (rated on a Likert scale).                                                                refinement of the question is needed (and will be undertaken


                                                                                                                                       IEEE A&E SYSTEMS MAGAZINE, JUNE 2008

       Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
Table 4. Handedness                                                      handedness. The results indicate that genuine signers and
                                                                                                 impostors did not rank signature within the same strength
                                                                                                 categories, that the impostors could not determine speed of
        Handedness                                        p-value                                the genuine signature, and that the impostors could not
                                                                                                 determine handedness. Furthermore, there were no common
                                                                                                 characteristics of the signature variables within the groups.
                                                                                                 Further research should be undertaken to examine whether
         Left                                               0.000                                these attributes change as the forger gains more knowledge
        Right                                               0.000                                about and experience with the signature.

                                                                                                 REFERENCES

in a subsequent study) to examine the PSS in more detail.                                               [I] D.J. Hamilton, J. Whelan, A. McLaren, 1. Maclntyre and A. Tizzard,
                                                                                                              Low cost dynamic signature verification system,
Table 2 outlines the results of the PSS category.
                                                                                                                   presented at 1995 European Convention on Security and
   When analyzing speeds (see Table 3, where 1 denotes fast                                                        Detection, Brighton, UK, 1995.
speed and 5 denotes slow speed), both groups had significant
differences, except for the neutral standing. Neither group                                                   Nelson and E. Kishon,
                                                                                                        [2] WA.
assigned a rank of 4 as a speed.                                                                             Use of dynamic features for signature verification,
   The results indicate that neither the impostors nor the                                                        presented at 1991 IEEE International Conference on Systems,
genuine users could determine the speed of each other's                                                           Man, and Cybernetics, Decision Aiding for Complex Systems,
                                                                                                                  Charlottesville, VA, 1991.
signatures. This is particularly interesting, as speed (or
velocity) is an often used a statistical feature in DSV
algorithms.                                                                                             [31   C. Vielhauer, R. Steinmetz and A. Mayerhofer,
                                                                                                                Biomnetric hash based on statistical features of online signatures,
    When analyzing handedness, the impostor group could not                                                          presented at 16' International Conference on Pattern
consistently determine the handedness of the genuine                                                                  Recognition, 2002.
signature owners. Only 3 out of 8. 37.5%, of the forgers
correctly identified a left-handed signature, while 3 of the                                            [4] F. Leclerc and R. Plamnondon,
left-handed signatures were not correctly identified at all.                                                   Automatic Signature Verification: The State of the Art - 1989-1993,
Comparably, 7 of 49 right-handed signatures were correctly                                                          Singapore: World Scientific Publishing Co., 1994.
identified by all forgers. However, the least accurate results
showed that 5 of 8 forgers incorrectly identified a signature                                           [5] J.-J. Brault and R. Plamiondon,
                                                                                                               A Complexity Measure of Handwritten Corves: Modeling
as left-handed when it was, in fact, right-handed. These
                                                                                                               of Dynamic Signature Forgery,
results are represented in Table 4                                                                                   IEEE Transactions on Systems, Man, and Cybernetics,
    The last question posed is whether the dynamic features                                                         Vol. 23, pp. 400-413, 1993.
extracted from the digitizer were similar for each group of
the PSS categories. For example: Do those in the                                                        [6] M.C. Fairhurst and S. Ng,
easy-to-forge category exhibit the same speed? Is there an                                                   Management of access through biometric control: A case study
underlying dynamic variable within these groups that are                                                     based on automatic signature verification,
                                                                                                                   Universal Access in the Information Society,
selected by impostors as easy to forge?                                                                            Vol. 1,pp. 31-39, 2001.
   An Analysis of Variance (ANOVA) test was conducted
over all of the individual variables that were extracted from                                           [71 A. Kholmatov and B. Yanikoglu,
the digitizer. At a = 0.05, none of these variables were                                                        Biometric Authentication Using Online Signatures,
significantly different across each difficultly group. For the                                                      presented at W International Symposium on Computer
                                                                                                                                   , 9

forger group, the ANOVA showed no significance with these                                                           and Information Sciences - ISCIS 2004,
extracted variables and difficulty group. There were some                                                           Kemer-Antalya, Turkey, 2004.
interesting correlations, however; speed was negatively
correlated with difficulty (-0.191, with p-value 0.273), as                                             [8] H.D. Crane and J.S. Ostem,
                                                                                                              Automatic Signature Verification using a Three-axis
were the number of strokes and difficulty of -0.314. The                                                      Force-Sensitive Pen,
forger groups had a slightly positive correlation with speed                                                       IEEE Transactions on Systems, Man, and Cybernetics,
and slightly negative correlation with segments (0.081).                                                           Vol. 13, pp. 329-337, 1983.

CONCLUSION                                                                                              [9] V.Nalwa,
                                                                                                                Automatic On-Line Signature Verification,
   The purpose was to assess whether genuine and impostor                                                           Proceedings of the IEEE, Vol. 85, pp. 215-239, 1997.
groups could successfully predict variables that could aid in
the successful forgery of the genuine signature. Variables                                              [10] M. Komiya Y.T.,
                                                                                                             On-line pen input signature verification PPI (pen-position/
included the perceived strength of the signature, speed, and

                                                                                                                                                                                      17
IEEE A&E SYSTEMS MAGAZINE, JUNE 2008
        Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
pen pressure / pen inclination),                                                                  [16] D-Y. Yeung. H. Chang, Y. Xiong, S. George, R. Kashi,
                presented at IEEE International Conference on Systems,                                            T. Matsumoto and G. Rigoll,
                Man, and Cybernetics, 1999,                                                                       SVC 2004: First International Signature Verification Competition,
                IEEE SMC '99 Conference Proceedings. 1999                                                              presented at First International Conference on Biometric
                Tokyo, Japan, 1999.                                                                                    Authentication, ICBA, Hong Kong, China, 2004.

     [I I] C. Schmidt and Kraiss, K-F.,                                                                     [17] S. Elliott,
            Establishment of Personalized Templates for Automatic                                                 A Comparison of On-Line Dynamic Signature Trait Variables
            Signature Verification,                                                                               vis-A-vis Mobile Computing Devices and Table-Based Digitizers,
                 presented at International Conference on Document                                                      in Third Workshop on Automatic Identification Advanced
                 Analysis and Recognition, 1997.                                                                        Technologies. Tarrytown, NY: IEEE, 2002.

     [12) W.S. Lee, Mohankrishnan, N. and Paulik, M.,                                                       [18] L. Lee, Berger, T. and Aviczer, E,
          Improved Segmentation through Dynamic Time Warping For                                                  Reliable On-Line Human Signature Verification Systems,
          Signature Verification using a Neural Network Classifier,                                                     IEEE Trans. Pattern Analysis Machine,
               presented at the 1998 International Conference on Image                                                  Vol. 18, pp. 643-647, 1996.
               Processing, 1998. ICIP 98. Proceedings, 1998.

                                                                                                            [19] W. Nelson and Kishon, F.,
     [13] Q.-Z. Wu, Jou, I-C. and Lee, S-Y,                                                                      Use of Dynamic Features for Signature Verification,
           On-Line Signature Verification using LPC Cepstrum and                                                      presented at Proceedings IEEE International Conference on
          Neural Networks,                                                                                             Systems, Man, and Cybernetics,
                IEEE Transactions on Systems, Man, and Cybernetics,                                                   Charlottesville, VA, 1991.
                pp. 148-153, 1997.

                                                                                                           [20] E.J.R. Justino, Bortolozzi, F. and Sabourin, R.,
     [141 M. Mingming and Wijesoma, W.,                                                                          The Interpersonal and Intrapersonal Variability Influences on
          Automatic On-Line Signature Verification Based on                                                      Off-Line Signature Verification using HMM,
          Multiple Models,                                                                                             presented at Proceedings of the XV Brazillian Symposium
               presented at Computational Intelligence in Financial                                                    on Computer Graphics and Image Processing,
               Engineering Conference, 2000.                                                                           (SIBGRAPHI'02), 2002.

     [15] P.-C.C. Chin-Chuan Han., Chao-Chih Hsu, and BorShenn Jeng,                                       (2 1] D. Cvrcek, Krhovjak, J. and V. Matyas,
           An on-line signature verification system using multi-template                                         PIN (& Chip) or signature -beating the cheating?
          matching approaches,                                                                                         Bmo, Czech Republic, 2005.
                presented at Security Technology, 1999 Proceedings,
                IEEE 33'ý Annual 1999 International Carnahan
                Conference, 1999.                                                                          [22] 1.Electronics, 2006,
                                                                                                                 E-Pad Signature Pad Specification Sheet.                             A




18                                                                                                                                    18
                                                                                                                                       IEEE A&E SYSTEMS MAGAZINE, JUNE 2008
         Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.

More Related Content

Similar to (2006) Dynamic Signature Forgery And Signature Strength

(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...
(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...
(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...International Center for Biometric Research
 
Efficient and secure authentication on remote server using stegno video objec...
Efficient and secure authentication on remote server using stegno video objec...Efficient and secure authentication on remote server using stegno video objec...
Efficient and secure authentication on remote server using stegno video objec...pradip patel
 
vargas2011.pdf
vargas2011.pdfvargas2011.pdf
vargas2011.pdfHouBou3
 
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docx
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docxRATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docx
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docxaudeleypearl
 
A Review on Robust identity verification using signature of a person
A Review on Robust identity verification using signature of a personA Review on Robust identity verification using signature of a person
A Review on Robust identity verification using signature of a personEditor IJMTER
 
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...ITIIIndustries
 
Fingerprint detection and face recognition for colonization control of fronti...
Fingerprint detection and face recognition for colonization control of fronti...Fingerprint detection and face recognition for colonization control of fronti...
Fingerprint detection and face recognition for colonization control of fronti...DR.P.S.JAGADEESH KUMAR
 
Comprehensive Review of Offline Signature Verification Mechanisms
Comprehensive Review of Offline Signature Verification MechanismsComprehensive Review of Offline Signature Verification Mechanisms
Comprehensive Review of Offline Signature Verification Mechanismsijtsrd
 
Behavioral biometrics mechanism for delaying password obsolescence
Behavioral biometrics   mechanism for delaying password obsolescenceBehavioral biometrics   mechanism for delaying password obsolescence
Behavioral biometrics mechanism for delaying password obsolescenceElaine Wooton
 
A comparative study of biometric authentication based on handwritten signatures
A comparative study of biometric authentication based on handwritten signaturesA comparative study of biometric authentication based on handwritten signatures
A comparative study of biometric authentication based on handwritten signatureseSAT Journals
 
A comparative study of biometric authentication based
A comparative study of biometric authentication basedA comparative study of biometric authentication based
A comparative study of biometric authentication basedeSAT Publishing House
 

Similar to (2006) Dynamic Signature Forgery And Signature Strength (20)

(2006) Perceived strength of signatures
(2006) Perceived strength of signatures(2006) Perceived strength of signatures
(2006) Perceived strength of signatures
 
(2007) The Challenges Associated with Laboratory-Based Distance Education
(2007) The Challenges Associated with Laboratory-Based Distance Education(2007) The Challenges Associated with Laboratory-Based Distance Education
(2007) The Challenges Associated with Laboratory-Based Distance Education
 
(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...
(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...
(2006) The Challenge of Forgeries and Perception of Dynamic Signature Verific...
 
Efficient and secure authentication on remote server using stegno video objec...
Efficient and secure authentication on remote server using stegno video objec...Efficient and secure authentication on remote server using stegno video objec...
Efficient and secure authentication on remote server using stegno video objec...
 
(2006) Keystroke Dynamics Verification Using a Spontaneously Generated Password
(2006) Keystroke Dynamics Verification Using a Spontaneously Generated Password(2006) Keystroke Dynamics Verification Using a Spontaneously Generated Password
(2006) Keystroke Dynamics Verification Using a Spontaneously Generated Password
 
Security Basics
Security BasicsSecurity Basics
Security Basics
 
Search Angels
Search AngelsSearch Angels
Search Angels
 
vargas2011.pdf
vargas2011.pdfvargas2011.pdf
vargas2011.pdf
 
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docx
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docxRATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docx
RATIO ANALYSIS RATIO ANALYSIS Note Please change the column names.docx
 
biometrics.ppt
biometrics.pptbiometrics.ppt
biometrics.ppt
 
A Review on Robust identity verification using signature of a person
A Review on Robust identity verification using signature of a personA Review on Robust identity verification using signature of a person
A Review on Robust identity verification using signature of a person
 
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...
Signature Forgery and the Forger – An Assessment of Influence on Handwritten ...
 
Kg2417521755
Kg2417521755Kg2417521755
Kg2417521755
 
Pdf1
Pdf1Pdf1
Pdf1
 
Fingerprint detection and face recognition for colonization control of fronti...
Fingerprint detection and face recognition for colonization control of fronti...Fingerprint detection and face recognition for colonization control of fronti...
Fingerprint detection and face recognition for colonization control of fronti...
 
Comprehensive Review of Offline Signature Verification Mechanisms
Comprehensive Review of Offline Signature Verification MechanismsComprehensive Review of Offline Signature Verification Mechanisms
Comprehensive Review of Offline Signature Verification Mechanisms
 
Behavioral biometrics mechanism for delaying password obsolescence
Behavioral biometrics   mechanism for delaying password obsolescenceBehavioral biometrics   mechanism for delaying password obsolescence
Behavioral biometrics mechanism for delaying password obsolescence
 
A comparative study of biometric authentication based on handwritten signatures
A comparative study of biometric authentication based on handwritten signaturesA comparative study of biometric authentication based on handwritten signatures
A comparative study of biometric authentication based on handwritten signatures
 
A comparative study of biometric authentication based
A comparative study of biometric authentication basedA comparative study of biometric authentication based
A comparative study of biometric authentication based
 
3d password - Report
3d password - Report  3d password - Report
3d password - Report
 

More from International Center for Biometric Research

An Investigation into Biometric Signature Capture Device Performance and User...
An Investigation into Biometric Signature Capture Device Performance and User...An Investigation into Biometric Signature Capture Device Performance and User...
An Investigation into Biometric Signature Capture Device Performance and User...International Center for Biometric Research
 
Advances in testing and evaluation using Human-Biometric sensor interaction m...
Advances in testing and evaluation using Human-Biometric sensor interaction m...Advances in testing and evaluation using Human-Biometric sensor interaction m...
Advances in testing and evaluation using Human-Biometric sensor interaction m...International Center for Biometric Research
 
(2010) Fingerprint recognition performance evaluation for mobile ID applications
(2010) Fingerprint recognition performance evaluation for mobile ID applications(2010) Fingerprint recognition performance evaluation for mobile ID applications
(2010) Fingerprint recognition performance evaluation for mobile ID applicationsInternational Center for Biometric Research
 

More from International Center for Biometric Research (20)

HBSI Automation Using the Kinect
HBSI Automation Using the KinectHBSI Automation Using the Kinect
HBSI Automation Using the Kinect
 
IT 34500
IT 34500IT 34500
IT 34500
 
An Investigation into Biometric Signature Capture Device Performance and User...
An Investigation into Biometric Signature Capture Device Performance and User...An Investigation into Biometric Signature Capture Device Performance and User...
An Investigation into Biometric Signature Capture Device Performance and User...
 
Entropy of Fingerprints
Entropy of FingerprintsEntropy of Fingerprints
Entropy of Fingerprints
 
Biometric and usability
Biometric and usabilityBiometric and usability
Biometric and usability
 
Examining Intra-Visit Iris Stability - Visit 4
Examining Intra-Visit Iris Stability - Visit 4Examining Intra-Visit Iris Stability - Visit 4
Examining Intra-Visit Iris Stability - Visit 4
 
Examining Intra-Visit Iris Stability - Visit 6
Examining Intra-Visit Iris Stability - Visit 6Examining Intra-Visit Iris Stability - Visit 6
Examining Intra-Visit Iris Stability - Visit 6
 
Examining Intra-Visit Iris Stability - Visit 2
Examining Intra-Visit Iris Stability - Visit 2Examining Intra-Visit Iris Stability - Visit 2
Examining Intra-Visit Iris Stability - Visit 2
 
Examining Intra-Visit Iris Stability - Visit 1
Examining Intra-Visit Iris Stability - Visit 1Examining Intra-Visit Iris Stability - Visit 1
Examining Intra-Visit Iris Stability - Visit 1
 
Examining Intra-Visit Iris Stability - Visit 3
Examining Intra-Visit Iris Stability - Visit 3Examining Intra-Visit Iris Stability - Visit 3
Examining Intra-Visit Iris Stability - Visit 3
 
Best Practices in Reporting Time Duration in Biometrics
Best Practices in Reporting Time Duration in BiometricsBest Practices in Reporting Time Duration in Biometrics
Best Practices in Reporting Time Duration in Biometrics
 
Examining Intra-Visit Iris Stability - Visit 5
Examining Intra-Visit Iris Stability - Visit 5Examining Intra-Visit Iris Stability - Visit 5
Examining Intra-Visit Iris Stability - Visit 5
 
Standards and Academia
Standards and AcademiaStandards and Academia
Standards and Academia
 
Interoperability and the Stability Score Index
Interoperability and the Stability Score IndexInteroperability and the Stability Score Index
Interoperability and the Stability Score Index
 
Advances in testing and evaluation using Human-Biometric sensor interaction m...
Advances in testing and evaluation using Human-Biometric sensor interaction m...Advances in testing and evaluation using Human-Biometric sensor interaction m...
Advances in testing and evaluation using Human-Biometric sensor interaction m...
 
Cerias talk on testing and evaluation
Cerias talk on testing and evaluationCerias talk on testing and evaluation
Cerias talk on testing and evaluation
 
IT 54500 overview
IT 54500 overviewIT 54500 overview
IT 54500 overview
 
Ben thesis slideshow
Ben thesis slideshowBen thesis slideshow
Ben thesis slideshow
 
(2010) Fingerprint recognition performance evaluation for mobile ID applications
(2010) Fingerprint recognition performance evaluation for mobile ID applications(2010) Fingerprint recognition performance evaluation for mobile ID applications
(2010) Fingerprint recognition performance evaluation for mobile ID applications
 
ICBR Databases
ICBR DatabasesICBR Databases
ICBR Databases
 

Recently uploaded

Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Mark Simos
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfAddepto
 
What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024Stephanie Beckett
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piececharlottematthew16
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticscarlostorres15106
 
SAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxSAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxNavinnSomaal
 
Training state-of-the-art general text embedding
Training state-of-the-art general text embeddingTraining state-of-the-art general text embedding
Training state-of-the-art general text embeddingZilliz
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr BaganFwdays
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxhariprasad279825
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clashcharlottematthew16
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostZilliz
 
My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024The Digital Insurer
 
Vector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesVector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesZilliz
 

Recently uploaded (20)

Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
Tampa BSides - Chef's Tour of Microsoft Security Adoption Framework (SAF)
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
Gen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdfGen AI in Business - Global Trends Report 2024.pdf
Gen AI in Business - Global Trends Report 2024.pdf
 
What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024What's New in Teams Calling, Meetings and Devices March 2024
What's New in Teams Calling, Meetings and Devices March 2024
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
Story boards and shot lists for my a level piece
Story boards and shot lists for my a level pieceStory boards and shot lists for my a level piece
Story boards and shot lists for my a level piece
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
SAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptxSAP Build Work Zone - Overview L2-L3.pptx
SAP Build Work Zone - Overview L2-L3.pptx
 
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptxE-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
E-Vehicle_Hacking_by_Parul Sharma_null_owasp.pptx
 
Training state-of-the-art general text embedding
Training state-of-the-art general text embeddingTraining state-of-the-art general text embedding
Training state-of-the-art general text embedding
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 
"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan"ML in Production",Oleksandr Bagan
"ML in Production",Oleksandr Bagan
 
DMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special EditionDMCC Future of Trade Web3 - Special Edition
DMCC Future of Trade Web3 - Special Edition
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptx
 
Powerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time ClashPowerpoint exploring the locations used in television show Time Clash
Powerpoint exploring the locations used in television show Time Clash
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
 
My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024My INSURER PTE LTD - Insurtech Innovation Award 2024
My INSURER PTE LTD - Insurtech Innovation Award 2024
 
Vector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector DatabasesVector Databases 101 - An introduction to the world of Vector Databases
Vector Databases 101 - An introduction to the world of Vector Databases
 

(2006) Dynamic Signature Forgery And Signature Strength

  • 1. Dynamic Signature Forgery and Signature Strength Perception Assessment Stephen Elliott & Adam Hunt Purdue University ABSTRACT include the fact that a signature is learned over time (and evolves over time as the owner and his or her handwriting Dynamic signature verification has many challenges matures), it contains variant measures (such as pressure, associated with the creation of the impostor dataset. The speed, etc., that can be changed), can be changed by the literature discusses several ways of determining the owner (depending on the ceremony of the transaction), and impostor signature provider, but this takes a different may have several versions (for example, at work and home approach - that of the opportunistic forger and his or her may have different signatures). relationship to the genuine signature holder. This A discussion of DSV invariably raises a number of examines the accuracy with which an opportunistic forger concerns. The first concern is that people acknowledge their assesses the various traits of the genuine signature, and failure to sign consistently, and the second is that most whether the genuine signature holder believes that his or people have attempted, irrespective of degree of success, to her signature is easy to forge. forge someone's signature at some time. In fact, a straw poll conducted in a class of 80 undergraduate college students INRODUCTION revealed that at least 90% of them have attempted to forge a signature at one time. When asked for more details about the Dynamic signature verification (DSV) has long been used forgery attempt, in the majority of cases, the subject of the to authenticate individuals based on their signing attack is someone who is known to the forger (typically a characteristics, such as speed, pressure, and graphical output. parent or close relative) and the signature is easily available. Approaches to DSV have been discussed in detail in the In these cases, it is more than likely that the forger has had literature. Popular applications, such as document several chances to practice the signature and that the authentication, financial transactions, and paper-based signature is not rigorously checked by the receiver of the transactions have all, at one time, used the signature to document being forged. These two conditions correlate to a convey the intent to complete a transaction [1, 2]. DSV is a low chance of getting caught - this is the scenario of the subset of a larger science called biometrics. Biometrics aims opportunistic forger. to authenticate an individual based on either behavioral or Another important consideration has to do with whether a physiological traits, (or a combination of both), including genuine signature holder believes that his or her signature is face recognition, iris recognition, and fingerprint recognition, difficult to forge, and whether the imposter also believes that to name a few. Many of these modalities are made up of both to be the case. The approach, proposed herein, is to behavioral and physiological attributes, with various understand whether the impostor can actually make proportions of each. Within the continuum, the signature is a well-informed decisions on the measurable variables of the strong behavioral biometric. The signature's unique traits genuine signature. For example: Can the forger determine the make it harder to test and evaluate than some of the other speed of the signature, as well as the handedness of the behavioral biometrics, such as voice or face recognition. genuine signer? If the forger can determine these most basic Challenges to testing and definitively evaluating the signature of attributes, then he or she might then achieve some level of success to forge some of the additional variables within DSV. VARIABLE CHARACTERISTICS OF THE DSV Author's Current Address: S. Elliott, Ph.D., and A. Hunt, Associate Professor, Department of Industrial Technology, College of Technology, Purdue University, 401 N. Grant Street, West Lafayette. IN 47906, DSV's numerous variables are calculated using the input USA. gathered from a digitizer. These variables include x and y Based on a presentation at Carnahsan 2006. (Cartesian) coordinates, pressure (p) or force, and time (t) 0885/8985/08/ USA $25.00 0 2008 IEEE [3]. This output from the digitizer is used to create the global IEEE A&E SYSTEMS MAGAZINE, JUNE 2008 1 13 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
  • 2. Table 1. DSV Studies & Error Rates Error Rates Study Authors Type I (FAR) Type 11 (FRR)EER Achemlal, Mourier, Lorette & Bonnefoy (1986) 11.0% 8.0% Beatson (1986) 1.0% 2.0% Bechet (1984) 5.0% 5.0% Bonnefoy & Lorette (1981) 0.6% Bault & Plamondon (1981) 1.2% 1.0% Collantier (1984) 3.5% Crane & Ostrem (1983) 1.5% 1.5% Debruyne (1985) 3.0% 2.0% Hale & Pagnini (1980) 1.5% 1.2/2.5% Herbst & Liu (1979) 1.7% 0.4% Herbst & Liu (1979) 2.4% 3.4% Ibrahim & Levrat (1979) 19.0% 5.5% Lam & Kamis (1979) 0.0% 2.5% Lorette (1983) 6.0% Mital, Hmn & Leng (1989) 0.0% 0.0% Parizeau & Ilamondon (1989) 4.0% Sato & Kogure (1982) 1.8% 0.0% Worthington, Chainer, Williford & Gundersen (1985) 1.8% 0.28-2.33% Zimmerman & Varady (1985) 30-50% 4-12% Cordelia, Foggia, Sanson & Vento 0.03-20.82% 5.7% Wirtz (1997) 10% Dimauro, Impedovo & Pirlo (1993) 1.7% 1.2% Naiwa (1997) 3% Naiwa (1997) 2% Nalwa (1997) 5% Mingming & Wijesoma. (2000) 5% Mingming & Wijesomna (2000) 7% Mingining & Wijesoma (2000) 9% Hamilton, Whelan, McLaren & Macintyre (1995) 34.0% 26.0% Hamilton, Whelan, McLaren & Macintyre (1995) 22.0% 18.0% Hamilton, Whelan, McLaren & Maclntyre (1995) 12.0% 10.0% Hamilton, Whelan, McLaren & Macintyre (1995) 7.0% 6.0% Martens & Clausen (1997) 1.5% 1.3% Chang, Wang & Suen (1993) 2.0% 2.5% Higashino (1992) 8.0% 0.6% Minot & Gentric (1992) 2.0% 4.0% Lucas & Damper (1990) 5.6% 4.5% Tseng & Huang (2002) 12.5-28.8% 5.0-12.5% Lee, Berger & Aviczer (1996) 1.0% 20.0% Lee, Berger & Aviczer (1996) 5.0% 20.0% Lee, Mohankrshnan & Paulik (1998) 0.9% 0.7% Han, Chang, Hsu & Jeng (1999) 4.0% 7.2% Komiya & Matsumoto (1999) 2% Cardot, Revenu, Victorri & Revillet (1993) 2.0% 4.0% Cardot, Revenu, Victorri & Revillet (1993) 0.9% 7.4% 14 IEEE A&E SYSTEMS MAGAZINE, JUNE 2008 14 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
  • 3. and local features described in various accounts [4, 51. These created by distorting real signatures through the addition of global and local features, derived from the basic set of data low-level noise and dilation or erosion of the various that a digitizer provides, vary significantly across algorithms. structures of the signature. [14] motivated the forgers by [6] outline 61 features, [7] note over 40 features, [8] discuss offering a cash reward. [15] examined people's signatures 44 features. In [2], describes the x and y coordinates of the over a four-month period to assess variability over time. In pen's motion. Also [9], observes that the temporal the Signature Verification Competition, genuine signers "characteristics the production of an on-line signatureare of created signatures other than their own [16]. In [17], the the key to the signature'sverification" (p. 5). From these author uses a number of different methodologies to generate many approaches and features, the number of variables the impostor distribution, with the majority of impostors associated with a dynamic signature can be synthesized to using some form of practice. In [ 18], the authors defined several major statistical features. These major statistical three different levels of forgeries: the simple, statically features include pressure, time, horizontal, and vertical skilled, and timed (p. 643). [19] used signatures that "on components of position, velocity, acceleration, and force, all casual visual inspection would pass as authentic" (p. 201). measured against time. An alternative approach to [20] provides three characteristics of forgery: the random characterizing a signature involves analysis of the "stroke," forgery, defined as one that belongs to a different writer of that is, the up and down movements of the pen on the the signature model; simple forgery, represented by a similar digitizer. The many dynamic traits collected by the digitizer shape consistency with the genuine signer's shape; and the during the act of signing are said to make an impostor skilled forgery (p. 2). signature easier to detect than that of a traditional paper-based impostor signature. PERCEIVED STRENGTH OF SIGNATURE (PSS) HIMPOSTOR SIGNATURE GENERATION The purpose of this paper is to assess the basic attack on a signature by an opportunistic forger and to determine the Impostor datasets are created in numerous ways and have perceived strength of the signature (PSS). PSS is a concept the effect of changing the respective performances of that indicates that an opportunistic forger will not forge a algorithms. This change can be done through the different signature that is difficult to forge, as their success at the generation of the impostor signatures. A review of the point-of-sale may be not as high as the forgery of an easy literature shows various performance results from several signature. This is more the trademark of an opportunistic studies, all of which have different methodologies for forgery than of a more sophisticated attack on the signature, collecting impostor signature datasets. Table 1 outlines the as outlined in previous research. For this study, an various studies and their respective error rates (false accept, opportunistic forger is analogous to an opportunistic thief, false reject, and the equal error rate where appropriate). that is, one who works on his or her own without any The variances in error rates shown in Table 1 (0% to 50% equipment [211]. This definition is further enhanced by the false accept rate and 0% to 20% false reject rate) can be absence of occasion to practice forging the signature. The explained by a number of factors, one of which has to do study outlines a basic truth involving the genuine signature with how an impostor signature dataset is created. [10]'s owner's perception of the strength of their signature and tries study is particularly interesting. This study had a database to understand whether the owner of a genuine signature has consisting of 293 genuine signatures and 540 forgery the same or different perception of the signature than that of signatures from eight individuals. Although the study did not the forger. explain how the forgery took place (in terms of training, In order to understand the basic truth of the perceived payment, etc.), eight individuals created the impostor dataset. strength of the signature, each of the genuine signature [11] study dataset was comprised of 496 original signatures owners was asked for the following information about their from 27 people. Each person signed I11 to 20 times. The signature: database contained 48 forgeries that "fulfill the requirement on the visual agreement and the dynamic similarity with the original signature" (p.5). [12] trained the algorithm using 1. How easy their signature was to forge (rated on 250 signatures per writer; of these 250 signatures, 100 were a Likert scale). authentic signatures and 150 were random forgeries, classified as the genuine signatures of other writers. [ 13] used 2. How fast or slow they signed their signature 27 people in their study, with the participants writing their (rated on a Likert scale from slow to fast). own signature. The study also used 4 people who imitated the signatures of these 27 people. Unfortunately, no further 3. Handedness (right- or left-handed; information is provided on the selection of the impostor or on ambidexterity was not an option captured by the what knowledge the forgers possessed in order to forge the survey). signatures. [1] used genuine signatures from other individuals as The objective in obtaining these three pieces of forgeries. In addition, a group of synthesized signatures was information was to assess whether the forger was able to IEEE A&E SYSTEMS MAGAZINE. JUNE 2008 15 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
  • 4. predict the speed and forgeability of genuine signatures, 2. How fast or slow was the genuine signature which was typically centered on the dynamic traits of speed estimated to be made (rated on a Likert scale or velocity, time, and graphical outline or complexity of the from slow to fast). shape. Furthermore, subjects were asked to sign their name on a digitizer so that feature variables could be extracted to 3. What was the handedness of the genuine subject estimate whether there were any correlations between (right- or left-handed; ambidexterity was not an variables and the respective PSS categories. option captured by the survey). Table 2. T-Test for Difficulty Groups The results were analyzed statistically to determine whether any significant differences existed between the genuine signature owners and the impostors regarding their assessments of the signature. RESULTS For the genuine dataset, a total of 60 subjects participated, of which 1 was female and 59 were males. Of these 60 subjects, 36 signed the digitizer. The remainder did not sign (or dropped out of the study). This represents a retention rate in the study of 60%. For the impostor dataset, there were 9 individuals who ranked the genuine signatures using the parameters previously described. A t-test was used to determine whether the mean of the METHODOLOGY genuine and impostor groups were statistically significant from each other with regards the three questions posed: the In order to assess PSS, two separate groups were rank of the perceived level of difficulty, velocity, and organized. The dataset of genuine signatures was collected handedness. A level of 0.05 was selected for determining from consent forms signed by the genuine signature owners. statistical significance. In the study, the data were normally The consent forms were used to maintain a level of distributed, and there were no outliers. The data were ranked ceremony, since the consent form is a document that requires from 1 to 5, with 1 being easy to forge and 5 being difficult a signature with a level of intent, as opposed to a random to forge. signature with no intent. This signature was subsequently used as the target signature. In order to estimate whether any Table 3. Signature Speeds of the dynamic signature verification variables were the same for each group (those who ranked their signature within the same Likert classification), each subject signed his or her Speed p-value name on a digitizer three times. In order to obtain a consistently precise signature, the study utilized an Interlink Electronics ePad-ink ProTM1 device, which has 100-400 reports 1 0.002 per second and 300 dots per inch [22]. The device was connected to forensic signature software 2 0.011 to extract the raw data from the digitizer, but the subjects 3 0.060 could not see the signature or the information on the PC monitor as they signed. The digitizer provided an inked No group 4 N/A display of the signature as the subject signed his or her name. 5 0.035 The three signatures were then processed and the resultant variables averaged across the signatures. The impostor group consisted of individuals other than When assessing the PSS, a t-test result shows parallels those who owned the original signatures. Members of the impostor group were asked for information about what they between genuine subjects and forgers when the level of difficulty was assessed as "neutral." Likewise, the difference observed while looking at the signed consent form of each in the "very difficult" ranking has a p-value of 0.053. Other individual in the genuine group: categories (levels 1, 2, and 4 in the Likert scale) exhibited significantly different means. It is difficult to determine 1. How easy the genuine signature was to forge whether the groups were statistically significant. Further (rated on a Likert scale). refinement of the question is needed (and will be undertaken IEEE A&E SYSTEMS MAGAZINE, JUNE 2008 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
  • 5. Table 4. Handedness handedness. The results indicate that genuine signers and impostors did not rank signature within the same strength categories, that the impostors could not determine speed of Handedness p-value the genuine signature, and that the impostors could not determine handedness. Furthermore, there were no common characteristics of the signature variables within the groups. Further research should be undertaken to examine whether Left 0.000 these attributes change as the forger gains more knowledge Right 0.000 about and experience with the signature. REFERENCES in a subsequent study) to examine the PSS in more detail. [I] D.J. Hamilton, J. Whelan, A. McLaren, 1. Maclntyre and A. Tizzard, Low cost dynamic signature verification system, Table 2 outlines the results of the PSS category. presented at 1995 European Convention on Security and When analyzing speeds (see Table 3, where 1 denotes fast Detection, Brighton, UK, 1995. speed and 5 denotes slow speed), both groups had significant differences, except for the neutral standing. Neither group Nelson and E. Kishon, [2] WA. assigned a rank of 4 as a speed. Use of dynamic features for signature verification, The results indicate that neither the impostors nor the presented at 1991 IEEE International Conference on Systems, genuine users could determine the speed of each other's Man, and Cybernetics, Decision Aiding for Complex Systems, Charlottesville, VA, 1991. signatures. This is particularly interesting, as speed (or velocity) is an often used a statistical feature in DSV algorithms. [31 C. Vielhauer, R. Steinmetz and A. Mayerhofer, Biomnetric hash based on statistical features of online signatures, When analyzing handedness, the impostor group could not presented at 16' International Conference on Pattern consistently determine the handedness of the genuine Recognition, 2002. signature owners. Only 3 out of 8. 37.5%, of the forgers correctly identified a left-handed signature, while 3 of the [4] F. Leclerc and R. Plamnondon, left-handed signatures were not correctly identified at all. Automatic Signature Verification: The State of the Art - 1989-1993, Comparably, 7 of 49 right-handed signatures were correctly Singapore: World Scientific Publishing Co., 1994. identified by all forgers. However, the least accurate results showed that 5 of 8 forgers incorrectly identified a signature [5] J.-J. Brault and R. Plamiondon, A Complexity Measure of Handwritten Corves: Modeling as left-handed when it was, in fact, right-handed. These of Dynamic Signature Forgery, results are represented in Table 4 IEEE Transactions on Systems, Man, and Cybernetics, The last question posed is whether the dynamic features Vol. 23, pp. 400-413, 1993. extracted from the digitizer were similar for each group of the PSS categories. For example: Do those in the [6] M.C. Fairhurst and S. Ng, easy-to-forge category exhibit the same speed? Is there an Management of access through biometric control: A case study underlying dynamic variable within these groups that are based on automatic signature verification, Universal Access in the Information Society, selected by impostors as easy to forge? Vol. 1,pp. 31-39, 2001. An Analysis of Variance (ANOVA) test was conducted over all of the individual variables that were extracted from [71 A. Kholmatov and B. Yanikoglu, the digitizer. At a = 0.05, none of these variables were Biometric Authentication Using Online Signatures, significantly different across each difficultly group. For the presented at W International Symposium on Computer , 9 forger group, the ANOVA showed no significance with these and Information Sciences - ISCIS 2004, extracted variables and difficulty group. There were some Kemer-Antalya, Turkey, 2004. interesting correlations, however; speed was negatively correlated with difficulty (-0.191, with p-value 0.273), as [8] H.D. Crane and J.S. Ostem, Automatic Signature Verification using a Three-axis were the number of strokes and difficulty of -0.314. The Force-Sensitive Pen, forger groups had a slightly positive correlation with speed IEEE Transactions on Systems, Man, and Cybernetics, and slightly negative correlation with segments (0.081). Vol. 13, pp. 329-337, 1983. CONCLUSION [9] V.Nalwa, Automatic On-Line Signature Verification, The purpose was to assess whether genuine and impostor Proceedings of the IEEE, Vol. 85, pp. 215-239, 1997. groups could successfully predict variables that could aid in the successful forgery of the genuine signature. Variables [10] M. Komiya Y.T., On-line pen input signature verification PPI (pen-position/ included the perceived strength of the signature, speed, and 17 IEEE A&E SYSTEMS MAGAZINE, JUNE 2008 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.
  • 6. pen pressure / pen inclination), [16] D-Y. Yeung. H. Chang, Y. Xiong, S. George, R. Kashi, presented at IEEE International Conference on Systems, T. Matsumoto and G. Rigoll, Man, and Cybernetics, 1999, SVC 2004: First International Signature Verification Competition, IEEE SMC '99 Conference Proceedings. 1999 presented at First International Conference on Biometric Tokyo, Japan, 1999. Authentication, ICBA, Hong Kong, China, 2004. [I I] C. Schmidt and Kraiss, K-F., [17] S. Elliott, Establishment of Personalized Templates for Automatic A Comparison of On-Line Dynamic Signature Trait Variables Signature Verification, vis-A-vis Mobile Computing Devices and Table-Based Digitizers, presented at International Conference on Document in Third Workshop on Automatic Identification Advanced Analysis and Recognition, 1997. Technologies. Tarrytown, NY: IEEE, 2002. [12) W.S. Lee, Mohankrishnan, N. and Paulik, M., [18] L. Lee, Berger, T. and Aviczer, E, Improved Segmentation through Dynamic Time Warping For Reliable On-Line Human Signature Verification Systems, Signature Verification using a Neural Network Classifier, IEEE Trans. Pattern Analysis Machine, presented at the 1998 International Conference on Image Vol. 18, pp. 643-647, 1996. Processing, 1998. ICIP 98. Proceedings, 1998. [19] W. Nelson and Kishon, F., [13] Q.-Z. Wu, Jou, I-C. and Lee, S-Y, Use of Dynamic Features for Signature Verification, On-Line Signature Verification using LPC Cepstrum and presented at Proceedings IEEE International Conference on Neural Networks, Systems, Man, and Cybernetics, IEEE Transactions on Systems, Man, and Cybernetics, Charlottesville, VA, 1991. pp. 148-153, 1997. [20] E.J.R. Justino, Bortolozzi, F. and Sabourin, R., [141 M. Mingming and Wijesoma, W., The Interpersonal and Intrapersonal Variability Influences on Automatic On-Line Signature Verification Based on Off-Line Signature Verification using HMM, Multiple Models, presented at Proceedings of the XV Brazillian Symposium presented at Computational Intelligence in Financial on Computer Graphics and Image Processing, Engineering Conference, 2000. (SIBGRAPHI'02), 2002. [15] P.-C.C. Chin-Chuan Han., Chao-Chih Hsu, and BorShenn Jeng, (2 1] D. Cvrcek, Krhovjak, J. and V. Matyas, An on-line signature verification system using multi-template PIN (& Chip) or signature -beating the cheating? matching approaches, Bmo, Czech Republic, 2005. presented at Security Technology, 1999 Proceedings, IEEE 33'ý Annual 1999 International Carnahan Conference, 1999. [22] 1.Electronics, 2006, E-Pad Signature Pad Specification Sheet. A 18 18 IEEE A&E SYSTEMS MAGAZINE, JUNE 2008 Authorized licensed use limited to: Purdue University. Downloaded on December 14, 2009 at 17:26 from IEEE Xplore. Restrictions apply.