Based on the Proposal I already made to make a presentation ppt,
prepare a 10-12 minute presentation. The content of the presentation should be the same as the project report (background, hypotheses, data explorations, discussion).
Need to do some analysis to get some chart …( This is a machine learning class, so need to use machine learning method)
And also with a speaking script for ppt.
427-project proposal
1. Title Page: Examining the Use of Touchscreen Biometrics for Smartphones
Authentication
2. Introduction:
In contemporary society, the cases of the loss of sensitive and private information
such as passwords, bank account details, social security numbers, emails, and
PINs in smartphones, have been on the rise. The rising cases of data loss, which
arise when smartphones are stolen or lost accidentally, can be attributed to a lack
of proper authentication systems. Data losses often expose mobile users to serious
data breaches and increasing the risk of cybercriminals accessing the sensitive
information to open fraudulent accounts, fund cybercrime activities, or using the
information to obtain medical care at the expense of legal users.
This research problem is of significant interest because it helps to unearth the
issue of data losses due to poor authentication system and propose the use of
touchscreen biometrics to safeguard sensitive and private information. Although
passwords, PINs, and mobile patterns have are often use to by smartphone users,
they can be compromised easily by unwarranted users, especially cybercriminals.
Currently, most of the authentication systems that are used by most smartphone
users are based on similar solutions leveraged when using laptops and desktop
computers, which often involve the application of a strong password, a PIN, or an
external security token device. However, these techniques are very cumbersome
to use, less secured, and prone to hacking. These issues often put users at serious
risks of data breaches and data losses.
Research Questions:
In this research, the key research questions are as follows:
1. Do smartphone biometrics offer a set of unique features that are unique to
each smartphone user?
2. Is it possible to authenticate smartphone users using constrained biometrics
on touchscreen-based smartphone devices?
3.
Which set of biometrics and behavioral features is the most discriminative?
4.
What privacy challenges can be addressed by using smartphone biometrics?
Challenges:
In this research, the major challenge is obtaining relevant datasets that can help in
understanding the relationship between smartphone touchscreen biometrics and
behavioral patterns of users when using their mobile devices. Hence, it will
require to invest in a huge amount of time, energy, and financial resources. Also,
available datasets have focused on conventional authentication systems, including
PINs, patterns, and passwords.
Literature Review:
Touchscreen biometric is gaining huge attention and many studies have been
conducted to examine its effectiveness in addressing the issue of mobile data
losses. Most of the existing studies agree that biometrics are much-needed
because provide the easiest way to identify legitimate users based on their eyes,
face, fingerprint identity, gesture, or swipe behavior. In a comparative study
conducted by Ngyuen & Voris (2017), where 10 smartphone users were sampled
for the study, they established that differences in behavioral patterns provide
distinguishing features that improve the effectiveness of touchscreen biometrics.
A similar study by Mecke et al. (2019), where 114 smartphone uses were used,
they revealed that smartphone biometrics are effective in enhancing data security
because users can modify and control typing features (hold time, flight time, or
touchscreen area). Mecke et al. (2019) also noted that biometrics do not leave
residues that could enable cybercriminals to decipher the secret touch patterns
Using 143,900 touchscreen operations of 71 participants, Shen et al. (2016)
established that touchscreen biometrics are effective data security schemes for
mobile devices. Leveraging classification techniques (support vector machine,
random technique, and nearest neighbor), they found that active authentication
activities enhance security depending on users’ characteristics such as angle
preferences, screen touch space, strength, and different rhythm of touchinteraction behavior. Fierrez et al. (2018) indicated that touchscreen swipe
biometrics and fingerprint biometrics that employs continuous authentication and
active authentication schemes offer a better security safeguard as compared to a
single-entry scheme.
3. Initial Hypotheses:
1. Biometrics used in smartphones provide higher data security because they
use users’ behavioral characteristics and touch-interaction behaviors to distinguish
users.
2. Touchscreen biometrics use active authentication schemes and continuous
authentication to provide better security safeguards as compared to single-entry
schemes used in conventional authentication strategies.
3. Smartphone biometrics as opposed to explicit authentication schemes
such as passwords and PINs offer unique features that prevent data losses.
4. Data-drivin Hypotheses:
The new data-driven hypotheses are as follows:
1. Integrating touchscreen biometrics into swipe dynamics, the password for
smartphones, and keystroke dynamics can enhance the security of data stored in
smartphones.
2. Touchscreen dynamic, including touch gestures and keystroking, and
accelerometer generated during human-mobile interaction provides an extra
security layer that protects data from being accessed by unwarranted users.
5. Proposed work and discussion:
Based on the datasets which contain gyroscope and accelerometer time-series
obtained from a smartwatch and smartphone as 51 test subjects perform 18
activities for 3 minutes each (UCI.edu, n.d), the proposed study uses datasets
(mobile phone’s gyroscope and accelerometer data), Equal Error Rate (EER),
False Acceptance Rate (FAR), and False Rejection Rate (FRR) to authenticate the
effectiveness of integrating touchscreen biometrics to traditional authentication
systems. The proposed study centers on the biometric effectiveness of activities,
including walking, writing, clapping, folding, et al.
The results are placed in the context of the literature to represent the human-phone
interaction (activities) and smartphone sensor signals used in smartphone
biometrics. Using the datasets, the study will try to identify the active
authentication schemas in the activities. The study intends to reveal that
touchscreen biometrics safeguard a mobile user’s private data because it leverages
continuous biometrics, which can be authenticated periodically by smartphone
owners to ensure higher security of their phones beyond the single-entry point. In
a nutshell, the study will try to build behavioral biometric models that depict the
effectiveness of touchscreen biometrics.
Dataset:
https://archive.ics.uci.edu/ml/datasets/WISDM+Smartphone+and+Smartwatch+Activi
ty+and+Biometrics+Dataset+#
References
Fierrez, J., Pozo, A., Martinez-Diaz, M., Galbally, J., & Morales, A. (2018).
Benchmarking touchscreen biometrics for mobile authentication. IEEE
Transactions on Information Forensics and Security, 13(11), 2720-2733.
Mecke, L., Buschek, D., Kiermeier, M., Prange, S., & Alt, F. (2019). Exploring
intentional behaviour modifications for password typing on mobile touchscreen
devices. In Fifteenth Symposium on Usable Privacy and Security ({SOUPS}
2019).
Ngyuen, T., & Voris, J. (2017). Touchscreen Biometrics Across Multiple Devices.
In SOUPS.
Shen, C., Zhang, Y., Guan, X., & Maxion, R. A. (2016). Performance analysis of touchinteraction behavior for active smartphone authentication. IEEE Transactions
on Information Forensics and Security, 11(3), 498-513.
UCI.edu. (n.d). WISDM Smartphone and Smartwatch Activity and Biometrics
Dataset Data Set. Retrieved October 13, 2020, from
https://archive.ics.uci.edu/ml/datasets/WISDM+Smartphone+and+Smartwatch+Activi
ty+and+Biometrics+Dataset+#
Project Proposal
HAIJIN YAO
10/16/2020
1. The Title Page:
Examining the Use of Touchscreen Biometrics for Smartphones Authentication
2. Introduction:
In contemporary society, the cases of the loss of sensitive and private information such as passwords, bank
account details, social security numbers, emails, and PINs in smartphones, have been on the rise. The
rising cases of data loss, which arise when smartphones are stolen or lost accidentally, can be attributed
to a lack of proper authentication systems. Data losses often expose mobile users to serious data breaches
and increasing the risk of cybercriminals accessing the sensitive information to open fraudulent accounts,
fund cybercrime activities, or using the information to obtain medical care at the expense of legal users.
This research problem is of significant interest because it helps to unearth the issue of data losses due to
poor authentication system and propose the use of touchscreen biometrics to safeguard sensitive and private
information. Although passwords, PINs, and mobile patterns have are often use to by smartphone users,
they can be compromised easily by unwarranted users, especially cybercriminals. Currently, most of the
authentication systems that are used by most smartphone users are based on similar solutions leveraged
when using laptops and desktop computers, which often involve the application of a strong password, a PIN,
or an external security token device. However, these techniques are very cumbersome to use, less secured,
and prone to hacking. These issues often put users at serious risks of data breaches and data losses.
####Research Questions: In this research, the key research questions are as follows: 1. Do smartphone
biometrics offer a set of unique features that are unique to each smartphone user? 2. Is it possible to
authenticate smartphone users using constrained biometrics on touchscreen-based smartphone devices? 3.
Which set of biometrics and behavioral features is the most discriminative? 4. What privacy challenges can
be addressed by using smartphone biometrics?
####Challenges: In this research, the major challenge is obtaining relevant datasets that can help in
understanding the relationship between smartphone touchscreen biometrics and behavioral patterns of users
when using their mobile devices. Hence, it will require to invest in a huge amount of time, energy, and
financial resources. Also, available datasets have focused on conventional authentication systems, including
PINs, patterns, and passwords.
####Literature Review: Touchscreen biometric is gaining huge attention and many studies have been
conducted to examine its effectiveness in addressing the issue of mobile data losses. Most of the existing
studies agree that biometrics are much-needed because provide the easiest way to identify legitimate users
based on their eyes, face, fingerprint identity, gesture, or swipe behavior. In a comparative study conducted
by Ngyuen & Voris (2017), where 10 smartphone users were sampled for the study, they established that
1
differences in behavioral patterns provide distinguishing features that improve the effectiveness of touchscreen biometrics. A similar study by Mecke et al. (2019), where 114 smartphone uses were used, they
revealed that smartphone biometrics are effective in enhancing data security because users can modify and
control typing features (hold time, flight time, or touchscreen area). Mecke et al. (2019) also noted that
biometrics do not leave residues that could enable cybercriminals to decipher the secret touch patterns Using
143,900 touchscreen operations of 71 participants, Shen et al. (2016) established that touchscreen biometrics
are effective data security schemes for mobile devices. Leveraging classification techniques (support vector
machine, random technique, and nearest neighbor), they found that active authentication activities enhance
security depending on users’ characteristics such as angle preferences, screen touch space, strength, and different rhythm of touch-interaction behavior. Fierrez et al. (2018) indicated that touchscreen swipe biometrics
and fingerprint biometrics that employs continuous authentication and active authentication schemes offer
a better security safeguard as compared to a single-entry scheme.
3. Initial Hypotheses:
1. Biometrics used in smartphones provide higher data security because they use users’ behavioral characteristics and touch-interaction behaviors to distinguish users.
2. Touchscreen biometrics use active authentication schemes and continuous authentication to provide
better security safeguards as compared to single-entry schemes used in conventional authentication
strategies.
3. Smartphone biometrics as opposed to explicit authentication schemes such as passwords and PINs offer
unique features that prevent data losses.
4. Data-drivin Hypotheses:
The new data-driven hypotheses are as follows:
1. Integrating touchscreen biometrics into swipe dynamics, the password for smartphones, and
keystroke dynamics can enhance the security of data stored in smartphones.
2. Touchscreen dynamic, including touch gestures and keystroking, and accelerometer generated during human-mobile interaction provides an extra security layer that protects data
from being accessed by unwarranted users.
5. Proposed work and discussion:
Based on the datasets which contain gyroscope and accelerometer time-series obtained from a smartwatch and
smartphone as 51 test subjects perform 18 activities for 3 minutes each (UCI.edu, n.d), the proposed study
uses datasets (mobile phone’s gyroscope and accelerometer data), Equal Error Rate (EER), False Acceptance
Rate (FAR), and False Rejection Rate (FRR) to authenticate the effectiveness of integrating touchscreen
biometrics to traditional authentication systems. The proposed study centers on the biometric effectiveness
of activities, including walking, writing, clapping, folding, et al. The results are placed in the context of
the literature to represent the human-phone interaction (activities) and smartphone sensor signals used in
smartphone biometrics. Using the datasets, the study will try to identify the active authentication schemas in
the activities. The study intends to reveal that touchscreen biometrics safeguard a mobile user’s private data
2
because it leverages continuous biometrics, which can be authenticated periodically by smartphone owners to
ensure higher security of their phones beyond the single-entry point. In a nutshell, the study will try to build
behavioral biometric models that depict the effectiveness of touchscreen biometrics. Dataset: https://archive.
ics.uci.edu/ml/datasets/WISDM+Smartphone+and+Smartwatch+Activity+and+Biometrics+Dataset+#
6. ** References:**
Fierrez, J., Pozo, A., Martinez-Diaz, M., Galbally, J., & Morales, A. (2018). Benchmarking
touchscreen biometrics for mobile authentication. IEEE Transactions on Information Forensics
and Security, 13(11), 2720-2733.
Mecke, L., Buschek, D., Kiermeier, M., Prange, S., & Alt, F. (2019). Exploring intentional
behaviour modifications for password typing on mobile touchscreen devices. In Fifteenth Symposium on Usable Privacy and Security ({SOUPS} 2019).
Ngyuen, T., & Voris, J. (2017). Touchscreen Biometrics Across Multiple Devices. In SOUPS.
Shen, C., Zhang, Y., Guan, X., & Maxion, R. A. (2016). Performance analysis of touchinteraction behavior for active smartphone authentication. IEEE Transactions on Information
Forensics and Security, 11(3), 498-513.
UCI.edu. (n.d). WISDM Smartphone and Smartwatch Activity and Biometrics Dataset Data
Set. Retrieved October 13, 2020, from https://archive.ics.uci.edu/ml/datasets/WISDM+
Smartphone+and+Smartwatch+Activity+and+Biometrics+Dataset+#
3
Purchase answer to see full
attachment
