IEEE Consumer Electronics Magazine - January/February 2022 - 31

EXPERIMENTAL RESULTS
To determine the validity of a " user review
process, " the acquired data were divided into
two sets: training data and test data. The training
data were collected in five sessions. Of which
three sessions were carried out to evaluate the
data. Using the test dataset, the classifier is
trained to evaluate predictive performance. This
study analyzed the two widely used performance
parameters for mobile-based behavior authentication
systems, called false rejection rate (FRR)
and the false accept rate (FAR). FRR treats authorized
users as unauthorized users and denies the
access (1), and FAR treats unauthorized users as
authorized users (2)
FRR ¼
No. of authorized user is denied access
No. of authorized users login attempts
(1)
FAR ¼
No. of unauthorized user access approvals
No. of unauthorized users login attempts
where Ci is the number of choices for position i,
in our case, which translate to
:
(2)
Total number of choices ¼
The obtained results show that the value of FAR
is 4.36 and FRR is 5.03.
SECURITY VERIFICATION
OF OP-GRID
The OP-Grid is a 7 11 matrix, with an image
at the background and numbers at the top of the
rows and columns. The headers of the rows and
columns change their positions randomly at
every login. This random position strengthens
the security as the pattern of the user will not be
at a specific row and column every time.
Shoulder Surfing and Smudge Attacks
The pattern drawn by the user on every login
session with small arithmetic operations of addition
and subtraction makes the login process of
OP-Grid very difficult for the shoulder surfer to
catch. The attacker cannot see the LP generated
by the server, as it disappears when the user
removes the finger from the screen. If attackers try
to login, then a new LP is generated; he will not be
able to perform the correct arithmetic operation.
Furthermore, if the attacker records the pattern
through the camera, still it will be an unsuccessful
attempt, because the LP and the headers of the
rows and columns change values randomly.
January/February 2022
Y4
i¼1
Ci:
(4)
After using fundamental combinatorics, this
quantity comes out to be 16 900. The same idea
can be applied to the authentication phase.
CONCLUSION
This article presents OP-Grid, a novel graphical
authentication scheme, which is the combination
of authentication factor " something you
process " and behavioral authentication. OP-Grid
is difficult to break as it has two lines of defense
against malicious user because on every login
session, the LP has a unique value, the changing
of the headers of rows and columns, and the
behavioral pattern of the user. This method can
further be strengthen by including physiological
and moving sensors for user identification.
&REFERENCES
1. S. S. ul Hasan Naqvi and S. Afzal, " Operation code
authentication preventing shoulder surfing attacks, " in
Proc. 3rd Int. Conf. Comput. Sci. Inf. Technol., vol. 4,
Jul. 2010, pp. 32-35.
2. L. Ginzberg, " User authentication system and
method, " US Patent 7,143,440, Nov. 28, 2006.
For a test set, a total of 650 samples were
used. An attacker has first to find the hidden values
of PA and PC. Assume that he knows the PA
and PC figures. Now, he has to draw the pattern
according to the user's behavior. Four attackers
have never been successful and some have been
successful in few tries. In a sitting posture, the
average rate of success of the attack was 3%.
The results showed that when an attacker draws
the OP-Grid pattern while walking, the shoulder
surfing attack and smudge attack are not possible.
Since, at the time of registration, the choices
of alphabets and digits are independent of each
other for n choices, it gives
Total number of choices ¼
Yn
i¼1
Ci
(3)
31
IEEE Consumer Electronics Magazine - January/February 2022

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - January/February 2022
