Fathima Institute of Engineering and Technology Adi Shankara

Fathima simna                                                    
                                prof 
Sundar .v

Department of Computer Science                                                   
 Department of Computer Science

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

Adi Shankara Institute of Engineering
and Technology                  Adi Shankara Institute of Engineering and

Technology

 

Abstract:Users are becoming more comfortable with
the idea of storing highly private information such as e-mails, photos, and
other sensitive documents on such devices. The popular mobile login methods
rely on numerical or graphical passwords. These techniques are vulnerable to
shoulder surfing attacks instigated by individuals with line-of-sight from a
short distance. Fingerprint is a popular biometric technique and has been used
for over 100 years 3 in different applications, including authentication on
mobile phones. But fingerprint authentication can fail if the fingerprint is
damaged. Electrocardiogram (ECG) methods have the advantage of concealing the
biometric features during authentication. However, complex hardware is required
to acquire this signal, making it hard to implement in mobile devices. In spite
of this, some companies have already developed ECG devices that work with
mobile phones 7 for medical monitoring purposes. To the best of our
knowledge, these devices have not been used for authentication in any
commercial product or research study.

 

IndexTerms—Biometrics,Electrocardiogram authentication,human identification

 

 

 

I INTRODUCTION

Users are more comfortable for storing highly
private information. popular login methods are numeric or graphic method. .
Fingerprint is the biometric method .if the fingerprint is damaged,
authentication fail . ECG method  have
the advantage of concealing the biometric features during authentication. Biometric technologies offer better security
mechanisms over traditional authentication methods  like 
password based ones, given the fact that the biometric feature is a
unique physiological characteristic that is always present and, depending on
the method used, may not be visible to other people. Electrocardiogram (ECG) methods
have the advantage of concealing the biometric features during authentication.
However, complex hardware is required to acquire this signal, making it hard to
implement in mobile devices

Traditional mobile login
methods, like numerical or graphical passwords, are vulnerable to passive
attacks. It is common for intruders to gain access to personal information of
their victims by watching them enter their passwords into their mobile screens
from a close proximity. With this in mind, a mobile biometric authentication
algorithm based on electrocardiogram (ECG) is proposed. With this algorithm,
the user will only need to touch two ECG electrodes (lead I) of the mobile
device to gain access. The algorithm was tested with a cellphone case heart
monitor in a controlled laboratory experiment at different times and conditions
with ten subjects and also with 73 records obtained from the Physionet
database. The obtained results reveal that our algorithm has 1.41% false
acceptance rate and 81.82% true acceptance rate with 4 s of signal acquisition.
To the best of our knowledge, this is the first approach on mobile
authentication that uses ECG biometric signals and it shows a promising future
for this technology. Nonetheless, further improvements are still needed to
optimize accuracy while maintaining a short acquisition time for authentication

 

II METHODOLOGY

ECG authentication algorithm that uses signals captured via a
lead I sensor . This allows users to input their biometric data by touching two
electrodes with their fingers. The employed ECG sensor is a practical mobile
phone heart monitor that produces a somewhat modest quality signal. Short time
limit for the length of the ECG authentication. Since the algorithm execution
time (order of milliseconds) is negligible compared with the signal capture
time, the overall authentication time must remain short.

fig 3.1. Preprocessing stages for the
ECG signal before applying the ECG authentication algorithm

 

An ECG signal En, shown in (1), with a length of
N, has to be treated before applying our authentication algorithm. The
steps involved in the enrollment and authentication are shown in Fig. 2. In the
first stage, peaks and valleys (fiducial points) from the ECG signal are
detected. This allows us to align and normalize the signal in order to avoid
the effects of changes in heart rate. Once the signal is normalized, we proceed
to extract the features. If we are enrolling a new user, the extracted features
allow us to create an enrollment template

to be stored in memory2. If we are authenticating a user, the extracted
features generate an authentication template. This authentication template is
used by the algorithm to authenticate a user against an enrollment template
(stored in memory)3. The following section explains these steps and also the
main contribution of this work, which is our ECG authentication algorithm:

Ee1, e2, . . . , eN?1,
eN .

 

 

Fig3.2.
 ECG
 authentication algorithm.

 

All the time-based features, except RTP (because R and
TP, which are used for normalization, are equivalent during enrollment and
authentication), are evaluated individually, and if they fall in the accepted
range of tolerance, a counter is increased. Once this process is completed, if
the value of the counter reaches an established score, the time-based features
are considered valid. Once the time-based features are validated, then the
algorithm proceeds to check the RS amplitude. If the RS amplitude falls into
the accepted range, the RQ amplitude will then be verified4. If it is within
an accepted threshold, then the system will validate the user. If any check of
these amplitude features fails, then the algorithm concludes that the
authentication template is a nonmatch. In this scheme, amplitude features are
given more prominence than the other features. If all the time-based feature
checks pass and the amplitudes checks do not, then a nonmatch conclusion is
reached.

The advantage of this hierarchical algorithm is that it takes
each feature individually. The outcome under the same conditions without the
hierarchical algorithm would be 0.21% FAR and 50.41% TAR. These results are
very impressive interms of FAR but not in terms of TAR, because a TAR of 50.41%
means that a genuine user will be rejected almost half of the attempts to
access the system. The algorithm was tested in three stages. The first stage is
to find the threshold values that will be used in the authentication
evaluation. The second stage is to assess the authentication accuracy of the.
The third stage is to assess the authentication accuracy of the algorithm
through an evaluation performed at the. In this paper, we will refer to the
term subject to the ECG data extracted from people at the laboratory and the
term record will refer to the data extracted from the Physionet database7. To
establish an appropriate time for authentication, we estimated the time
required for the users to input their traditional passwords in their phones. It
was found that it takes them an average time of 4 s from ten users in the
laboratory.

 

 

III
CONCLUSION

ECG Authentication approach is for mobile phones.
Proposed algorithm uses a heirarchical scheme that reduces the acquition time.
Proposed algorithm is reliable as other exists.81.82%
TAR and a 1.41% FAR. Advantage of a supporting shorter acquisition  time (4 s).

Our future works will aim to improve the TAR and FAR using
machine learning algorithm. The obtained results show
that the algorithm is suitable to Work with mobiles and with other sensors; as
it was tested also With Physionet database. To the best of our knowledge, this
Is the first ECG authentication approach designed exclusively For mobile
phones.

The obtained results reveal that our algorithm has
1.41% false acceptance rate and 81.82% true acceptance rate with 4 s of signal
acquisition. To the best of our knowledge, this is the first approach on mobile
authentication that uses ECG biometric signals and it shows a promising future
for this technology.

IV REFERENCES

1   J. S.
Arteaga-Falconi, H. A. O. and Saddik, A. E. (2016). Ecg authentication for
mobile devices. In IEEE Transactions on Instrumentation and Measurement, pages
591–600. J. S. Arteaga-Falconi and Saddik, 2016

2   G.
D. Clark and J. Lindqvist, “Engineering gesture-based authentication systems,” IEEE Pervasive Comput., vol. 14,
no. 1, pp. 18–25, Jan./Mar. 2015.

3   E.
H. Holder, Jr., L. O. Robinson, and J. H. Laub, “The fingerprint sourcebook,” Dept. Justice, Office Justice Programs,
Nat. Inst. Justice, Office Justice Programs, Washington,
DC, USA,

4   P.
J. Phillips, A. Martin, C. L. Wilson, and M. Przybocki, “An introduction
evaluating biometric systems,” Computer, vol. 33, no. 2, pp. 56–63

5  
M. Espinoza, C. Champod, and
P. Margot, “Vulnerabilities of fingerprint reader
to fake fingerprints attacks,” Forensic Sci. Int., vol. 204, nos. 1–3, pp. 41–49, Jan. 2011.

6  
(2013). Chaos Computer
Club breaks Apple TouchID. Online. Available:
http://www.ccc.de/en/updates/2013/ccc-breaks-apple-touchid

7  
Heart Monitor AC-002 User
Manual,
AliveCor, Inc., San Francisco, CA, USA,
Apr. 2015