Three Level Face Recognition and Pattern Lock

Authentication System
AJAY A. JAISWAL1, ADITYA MANWATKAR 2, SANSKRUTI KUMAR 3, SUDHANSHU KHODAKE 4,
SHRUTI DHOTE 5, SEJAL CHIDDARWAR 6, NARAYANI FOPSE 7
1,2,3,4,5,6,7
Department of Artificial Intelligence and Data Science, K. D. K. College of Engineering,
Nagpur, Maharashtra, India
1
[email protected]
2
[email protected]
3
[email protected]
4
[email protected]
5
[email protected]
6
[email protected]
7
[email protected]
Abstract— The most commonly used type of password for system detects the face using camera and at last, the user can
security purposes is text-based. However, these passwords can be add any pattern lock for its final level Authentications. The
easily compromised, putting personal data at risk. With the rise main objective of this project is to improve system security
in cybercrime, login and access security threats have become a based on extensive research. Three levels of authentication
major concern. To enhance security, we’ve developed a Three- offer more robust protection compared to one or two levels.
Level Password Authentication system. This system includes Users enter critical details and log in securely across all three
three levels of logins, each with a different type of password:
levels.
Passphrase, Face Recognition, and Pattern lock. As users
progress through the levels, the password complexity increases,
ensuring more secure access. This Web-based Three-Level
Authentication System helps users safeguard their data from
hackers and cyber threats.

Keywords— Authentication, Face Recognition, Pattern Lock,

User Verification, 3-level security

I. INTRODUCTION
In today’s digital landscape, ensuring robust security for
user data and system access is paramount. Traditional text-
based passwords, while widely used, are susceptible to
breaches. As cybercrime continues to evolve, novel
authentication methods are essential to safeguard sensitive Fig. 1.1 Registration Process
information.
Our proposed Three-Level Authentication
System addresses these concerns by combining three distinct
layers of security. During user registration, individuals set up
a three-tiered password system:
1. Conventional Password (Level 1):
 Users create a text-based password during
registration.
 This level serves as the baseline authentication
method.
 The system verifies the entered password against
stored data in the database.

2. Facial Recognition (Level 2):

 Leveraging advanced biometric technology, users
associate their face with their account.
 During login, the system captures and analyzes facial
features.
 Successful recognition grants access to the next level.
3. Graphical Password (Level 3):
 Users define a unique pattern (e.g., drawing on a Fig. 1.2 Login Process
grid) as their third-level password.
 The pattern lock adds an additional layer of
complexity. II. LITERATURE REVIEW
 Successful pattern matching completes the Several papers were reviewed, each offering unique
authentication process. perspectives on implementing effective authentication
methods for security. Here’s a summary:
Our proposed Three-Level Face Recognition and Pattern [1] In June 2017, by Karan Maheshwari and Nalini N
Lock Authentication System significantly enhances security proposed a “Facial Recognition Enabled Smart Door Using
while maintaining user-friendliness. By incorporating three Microsoft Face API”. The paper discusses a smart door
distinct levels of authentication, we address vulnerabilities system that uses Microsoft’s Face API for facial recognition.
such as Man-in-the-Middle attacks and Brute-force attacks. A It identifies individuals via a live HD camera, provides voice
three-level security system is a time-consuming approach outputs, and uses electromagnets for door locking. The system
since the user needs to enter details carefully at first level, and is hosted on Microsoft Azure cloud support.
second level can only be passes by the user itself because the
 [2] In 2018, Aparna M and Anjusree C M introduced 1) Level 1: Textual Authentication: In the initial phase
a “Three-level security system using Image-based of authentication, users log in using a combination of
Authentication”. This paper proposed the use of a One-Time a UserID and a Password. The password must be at
Password (OTP) as the third level of security. They suggested least 11 characters long and can include
an image choice authentication, where users can select a alphanumeric characters and symbols. If a user
specific image from given options as the second level. The forgets their password or enters an invalid one, they
authors proposed various types of authentication systems, all have the option to use their registered email address
of which offer high security. for verification. After confirming their username via
[3] In April 2018, B. Lakshmi Praveena, M. Anitha, email, users can proceed with the three-level
J. Supriya, and T. Lakshmi Priya proposed a “Student Portal password authentication process.
with 3 Level Passwords Authentication System”. This project
included a text password (passphrase), a colour password, and
a pattern password for the three respective levels.
[4] In June 2020, Rahul Chourasia proposed a
“Three-level password authentication system”. This paper
suggested a novel approach for textual passwords. They
recommended replacing textual passwords with graphical
passwords, which are easier to remember and more user-
friendly. Moreover, graphical passwords offer enhanced
security.
[5] In July 2022, N Chaitra, Pratibha, and Dr.
Rajashree V Biradar proposed a “Three-level password
authentication system”. According to this paper, the three-
level authentication is a combination of three existing
schemes: text-based password, pattern lock, and a one-time
password (OTP).

III. PROPOSED SYSTEM

Fig. 3.1 Level 1: Textual Authentication
In the digital world, safeguarding our valuable assets relies 1) Level 2: Face Recognition: During user registration,
heavily on authentication methods. While we commonly use individuals provide facial images, which serve as
text-based passwords, they have limitations. Biometric reference points for subsequent authentication. When a
techniques, though powerful, can be computationally user attempts to log in, the system captures their live
demanding. But passwords are more than just keys—they facial image using a webcam or camera. This live
safeguard our personal identity, ensuring privacy and image is then compared with stored reference images
protecting sensitive data. Additionally, once we authenticate a in the system’s database. If the facial features match
transaction using a password, we cannot later deny its validity. any of the enrolled faces, then ask for 6-digit pin. If 6-
To address these challenges, we propose the Three-Level digit pin matches with user security pin, access to the
Face Recognition and Pattern Lock Authentication next level is granted. Facial Recognition is done
System. By combining text, facial recognition, and graphical through FaceIO API which is open-source platform for
patterns, we enhance security and prevents cyber threats. Face recognition.
This paper proposes Three levels of security are as below:

Fig. 3.2 Level 2: Face Recognition

2) Level 3: Pattern Lock Authentication: In our

innovative Three-Level Face Recognition and Pattern
Lock Authentication System, the third level introduces
the familiar Android lock pattern. Users create this
Fig. 2.1 Proposed Authentication System pattern by connecting dots on a grid, resulting in a
unique sequence. During registration, individuals
 choose a personalized pattern by connecting specific comprehensive defence against unauthorized access,
dots on the grid. These selected dots serve as their prioritizing safeguarding sensitive information. As technology
reference for subsequent logins. When a user attempts evolves, continuous improvements in security protocols are
to log in, they recreate their chosen pattern by essential.
connecting the dots in the same order. Successful
pattern matching grants access to the system.
V. FUTURE SCOPE
The future work plan is centred around enhancing both the
security and user experience within our system, with a
primary focus on four key project areas. In the domain of
biometric authentication, the strategy involves a thorough
examination of advanced techniques such as iris recognition,
fingerprint scanning, and vein patterns. Additionally, there
will be an exploration of multi-modal biometrics,
incorporating face, voice, and other features to strengthen
authentication. A crucial aspect of this exploration entails
assessing the trade-offs among accuracy, usability, and
privacy for each biometric method.

Extending our existing three-level authentication system to

include a fourth level represents a significant initiative. This
Fig. 3.3 Level 3: Pattern Lock authentication extension involves integrating voice or speech recognition as
the fourth layer, offering unique vocal patterns for
authentication. To mitigate potential impersonation attacks,
The system comprises three login stages, each with a
we will implement robust anti-spoofing measures, subjecting
distinct password mechanism. The complexity of the
them to comprehensive testing to ensure effectiveness. The
password escalates with each successive stage. Users are
implementation process will be meticulously documented,
required to input the correct password for successful login and
providing clear guidelines for configuration and optimization.
have the freedom to establish their own passwords. Even if a
programmer manages to breach the first security level, they
To ensure real-time tracking of user behaviour and swift
would be unable to bypass the second level, which involves
identification of suspicious activities, we will develop
face recognition, and the third level, which requires the user to
continuous monitoring and anomaly detection mechanisms.
draw a unique grid-based pattern lock.
An alert system will be established to promptly notify
administrators or users of unusual login patterns or potential
This paper introduces a three-tier authentication system. security threats, with defined escalation procedures for
New users must register for all three levels. Once registered, varying alert levels.
users can log in. However, if an incorrect password is entered,
login is denied. A password recovery option is available, Central to our objectives is the enhancement of user
where users can verify their username with their email ID to experience, which involves collecting user feedback to
change their password. All user data is stored in a database, as identify pain points and areas for improvement. This valuable
illustrated in the subsequent diagram. data will guide the optimization of the user interface, aiming
for seamless interaction across all authentication levels.
IV. CONCLUSION Throughout the design and implementation process, we will
In our research, we aimed to enhance system security by prioritize usability, accessibility, and responsiveness. The
exploring various authentication methods. Through an overarching goal of the work plan is to strike a harmonious
extensive survey of published papers, we found that three- balance between heightened security measures, the
level authentication systems offer superior security compared incorporation of advanced authentication methods, and an
to one or two-level approaches. Our system incorporates three overall improvement in user experience. Regular feedback
distinct authentication levels, each adding complexity. Users loops, usability testing, and iterative design processes will be
must navigate through these phases, ensuring a robust security integral to ensuring the continuous evolution of the system to
framework. We begin with a basic text-based user ID and meet both security and user satisfaction requirements.
password authentication as the first level. The second level
introduces Face Recognition and third as Grid based Pattern REFERENCES
Lock authentication (often refer to as Android pattern lock). [1] Karan Maheshwari, Nalini N, “Facial Recognition Enabled Smart Door
Using Microsoft Face API”, International Journal of Engineering
While time complexity may increase due to multi-level Trends and Applications (IJETA) – Volume 4 Issue 3, May-Jun 2017
authentication, data security remains paramount. In
conclusion, our three-level authentication system provides a
[2] M. Aparna, S. Gopalakrishnan, C. M. Anjusree, “Three Level Security [6] Gouri Sankar Mishra, Pradeep Kumar Mishra, Parma Nand, Rani
System using Image Based Authentication”, IJARCCE, Vol. 7, Issue Astya, Amrita, “User Authentication: A Three Level Password
11, November 2018 Authentication Mechanism” Conference Series, Volume 1712,
[3] B. LAKSHMI PRAVEENA, M. ANITHA, J. SUPRIYA, T. International Conference On Computational Physics in Emerging
LAKSHMI PRIYA, “Student Portal with 3 Level Passwords Technologies 1 August, Mangalore, India, 2020.
Authentication System”, © APR 2018 | IRE Journals | Volume 1 Issue [7] https://ieeexplore.ieee.org/abstract/document/9137776/
10 | ISSN: 2456-8880 [8] https://ieeexplore.ieee.org/iel7/6287639/6514899/08721054.pdf
[4] RAHUL CHOURASIA, Dr. N. PARTHEEBAN, “Three level [9] https://aws.amazon.com/what-is/facial-recognition/
password authentication system” © 2020 IJCRT | Volume 8, Issue 6 [10] https://en.wikipedia.org/wiki/Facial_recognition_system
June 2020 | ISSN: 2320-2882 [11] https://faceio.net/
[5] N Chaitra, Pratibha, Dr. Rajashree V Biradar, “THREE LEVEL
PASSWORD AUTHENTICATION SYSTEM” IRJMETS,
Volume:04/Issue:07/July-2022 Impact Factor- 6.752