PIMPRI CHINCHWAD EDUCATION TRUST'S

PIMPRI CHINCHWAD COLLEGE OF ENGINEERING

FINAL SEMINAR REVIEW

INFORMATION SECURITY
Topic:
Critical Segment Based Real-time ESignature for Securing Mobile
Transactions

GUIDE:Mrs. Reena Kharat

PRESENTED BY:ANIRUDDHA SHINDE

PROBLEM STATEMENT
Rising use of smartphones has seen a rapid rise since last few years
in the field of e-commerce, secure transaction, approving financial
transactions, online voting system. This rise has led to several
challenges in terms of security and confidentiality. To ensure
successful deployment of transaction system which is secure at your
fingertips requires a strong verification system

KEY FEATURES(of solution):l

E-Signature(Stylus, Single Finger, Multi Finger Input)
l
Signature verification leveraging Multitouch Environment
l
Real-Time Verification
l
Robust to signature forging attacks

NEED FOR E-SIGNATURE

User signature verification becomes critical to ensure the success
deployment of online transactions such as approving legal documents
and authenticating financial transactions.
E-Signature exploits the most basic touch platforms which are common
in all the smartphones(low end or high end).
It's biggest advantage being the utilization of machine learning is
possible.
It is a real time application.

E-Signature
Now what exactly is an E-Signature????
l
An electronic signature, or e-signature, is any electronic means that
indicates either that a person adopts the contents of an electronic
message, or more broadly that the person who claims to have written a
message is the one who wrote it.
l
When implemented in compliance to digital signature standards, e-signing
should offer end-to-end privacy with the signing process being user-friendly
and secure.
l One of the best examples of widespread use of electronic signature is in
the Baltics. Estonia, and most recently Latvia, have seen widespread
adoption of E-Sign systems by which all citizens are provided with the
means to electronically sign agreements
l Many companies have been aiming to launch applications for mobile phones
so that people can sign documents electronically when viewing contracts or
other documents on mobiles.
l eSign is an online electronic signature service in India to facilitate an
Aadhaar holder to digitally sign a document.

E-Signature:How they work!!!!

There are typically two types of e-signature system:
Static: In this mode, users write their signature on paper, digitize it through an
optical scanner or a camera, and the biometric system recognizes the
signature analyzing its shape. This group is also known as off-line.
Dynamic: In this mode, users write their signature in a digitizing tablet, which
acquires the signature in real time. Another possibility is the acquisition by
means of stylus-operated PDA
l

Currently used Systems....

l

l
l
l

Currently used systems basically use some sort of device to incorporate a digitized signature
devices like stylus or Biometric smart pen(BiSP).
In BiSP system:
The dynamic of a signature is captured by sensors mounted inside the pen.
The recorded signature signals (i.e., signature time series) are presented to the authentication
system
Where they are classified by different static and dynamic classifiers using different features

BiSP System
l

If a person shall be verified by his/her signature,

i.e., the identity of this person shall be confirmed,
the following steps have to be conducted:
1)The person who wants to be authenticated
mustprovide a registered ID, such as his/her
name, and a signature recorded with the BiSP.
2) The person-specific reference model for the
provided ID, which has been generated during the
enrollment, is loaded from a database.
3) The offset of the recorded signature is
eliminated.Additionally, the signature is trimmed.

Trimming is the process in which synchronisation

of recording start and end is acquired.
A more stable sample is achieved by eliminating
signature offset.

Suggested System
Proposed system consists of three main components:Critical Segment
Extraction,Feature Extraction and Signature Normalization and Interpolation.
Critical Segment Extraction is the core component that captures the intrinsic
user signing behaviors by identifying the segments which remain stable within
the users genuine signature.
Feature Extraction in this our system extracts useful features to describe both
the geometric layout of the signature as well as the users signing behavior by
leveraging the rich set of information enabled by touch screens.
Signature Normalization and Interpolation is performed to reduce the impact
of signature geometric distortions caused by different writing sizes and
orientations on touch screens.
System identifies the users signing behavior and exploits the geometric
layout of the signature and the users behavior and physiological
characteristics to achieve accurate and robust online signature verification.

Algorithms used
The proposed system mainly makes extensive use of one algorithm
which is used in the critical segment identification of the signature.
The algorithm is referred to as Critical Segment Identification
Algorithm.
The above mentioned algorithm works functions by broadly
distributing the identification process into two well known
algorithms:
1.DTW-Dynamic Time Warping
2.DMS-Direct Matching Sample

Critical Segment Identification Algorithm

A signature can be decomposed into several stroke segments.
However, only a few of these decomposed segments are invariant
across a set of signatures a user signs.
Such segments reflect the users intrinsic signing behaviour, and we
refer to them as critical segments.
To identify the critical segments the algorithm takes a pair of
signatures from the user as inputs and compare them using the
dynamic time warping (DTW).
Dynamic time warping(DTW) is analgorithmfor measuring
similarity between two temporal sequences which may vary in time
or speed.

Working of DTW & DMS

Given a feature sequence used to represent the signature (e.g., x and
y coordinates of the signature, the signing pressure of the signature),
the resulting coupling sequence from DTW denotes an optimal
alignment between two feature sequences.
Direct matching samples (DMSs) in the coupling sequence represent
the segments without significant distortion between the two input
signatures. Thus, the DMSs extracted from the coupling sequence
can be utilized to derive a weight vector which denotes the similarity
between two signatures.

Critical Segment Identification Algorithm

Capture the invariance b/w users genuine signatures in a signature
pool, and then average over all the weight vector to extract the
critical segments as these segments have high similarity among a
group of genuine signature.
To simplify the description of the critical segment extraction
algorithm, we assume a signature is already normalized and
interpolated.
A direct matching sample (DMS) in the coupling sequence is defined
as a feature sample in the q1- th signature which has an one-to-one
coupling with a sample in the q2-th feature sequence.
W=1 if DMS found ,0 otherwise (1)

Critical Segment Identification Algorithm

Figure 2 (a) plots the pressure
features extracted from two
signatures. From this figure, we
observe that the first 4 samples are
very similar, while the rest samples
between two signatures differ
significantly. We then generate the
coupling sequence between two
features based on DTW.
In Figure 2 (b), the direct matching
samples (DMSs) in the coupling
sequence are shown as solid dots
and the non-direct matching samples
are shown as hollow dots. Use
Equation (1).The Algorithm
successfully helps to extract 4
critical segments from a signature.

Feature Extraction Segment

Every touch event is characterized by the following information: a finger ID
uniquely assigned to each finger, x and y coordinates of the touch point,
pressure, and the time stamp of the event. We thus can use such information
as the basis to describe the characteristic of signature that consists of a
series of touch events.
Based on the information of each touch event, our signature verification
system extracts several features to describe both the geometric layout of the
signature as well as the users signing behaviour.

Feature Extraction Signature Normalization and

interpolation
Common features are distance b/w fingers, x and y co-ordinates, x and y
velocity, touch pressure, etc. After feature extraction user profile is
constructed and matched with the pre-constructed signature profile for
verification.
Signature normalization is done to provide robust system independent of
touch screen size, font, orientation, etc. Its achieved via normal translation,
scaling and rotation operations.
The user usually writes signatures at different speeds which produce different
numbers of touch events of signature. To deal with variable number of touch
events, our system further performs signature interpolation. This step allows
us to perform robust signature verification by directly measuring the similarity
between two identical length of touch events.

Overview of system(all components together)

Motivation.....
l
l

l
l
l
l

Society aspects:
Provides easy and secure transactions which inturn finds
applications in online voting system and other authentication based
systems(signing legal documents).
Environment aspects:
Eco-Friendly as use of papers is completely avoided.
Technology aspects:
Easy,reliable and robust with requirement of only a touch screen
phone.

Expected Outcome......
The proposed system identifies the critical segments, which remain invariant
within a users signature, to capture the users intrinsic signing behavior.
By leveraging the rich set of information enabled by touch screens, our
system extracts useful features to describe both the geometric layout of the
signature as well as a users behavioral and physiological characteristics
during the signing process.
Five or more features result in acceptable performance with over 95% true
positive rate and less than 5% false positive rate in all these three scenarios.
Furthermore, similar true positive rate and false positive rate are achieved
even if the attacker performs imitation attacks

References....
1)Yanzhi Ren, Chen Wang, Yingying Chen, Mooi Choo Chuah, Jie Yang Critical
Segment Based Real-time E-Signature for Securing Mobile Transactions 2015
IEEE Conference on Communications and Network Security (CNS).
2)Sunitha N.R., Amberker B.B A New method of verifying Digital Signatures in
E- Check Processing ICON 2008 16 th IEEE International Conference.
3)Christian Gruber Christian Hook,Jurgen Kempf,Georg Scharfenberg,and
Bernhard SickA Flexible Architecture for Online Signature Verification Based
on a Novel Biometric Pen2006 IEEE
-Wikipedia
-wired.com

THANK YOU...........