Facial Recognition Door unlock using Machine Learning.

Facial Recognition Door Unlock
Submitted in the partial fulfilment
for the award of the degree of
BACHELOROF ENGINEERING
IN
Computer Science and Engineering
with Specialization in
Artificial Intelligence and Machine Learning
Submitted by:
Sagar Vashist 21BCS6597
Yuboraj Ganguly 21BCS6602
Sameesh Bagga 21BCS6623
Under the Supervision of:
Mrs. Rajat Tiwari

With the rise of smart home technology, many are looking for ways
to increase the security and convenience of their homes & property.
One potential solution is a facial recognition door unlock system,
which uses advanced technology to grant access to authorized users.
Facial recognition technology has become increasingly sophisticated
in recent years, thanks to advances in artificial intelligence and
machine learning. These systems can now accurately recognize
individuals based on facial features, even in challenging lighting
conditions or from unusual angles.
In addition to providing a high level of security, a facial recognition
door unlock system can also be very convenient for users. They don't
need to remember a code or carry a key; instead, they can simply
approach the door and be granted access automatically.
Introduction

This project aims to explore the potential of facial recognition
technology for home security, by building a prototype facial
recognition door unlock system. We will use a combination of
hardware and software tools to create a system that can accurately
recognize authorized users and grant them access to a locked door.
While facial recognition technology has many potential benefits, it's
also important to consider the potential risks and challenges. We will
take a comprehensive approach to security and privacy, using
advanced encryption and ensuring that user data is protected at all
times.
Ultimately, the goal of this project is to create a secure and
convenient facial recognition door unlock system that can be used in
real-world applications. By demonstrating the potential of this
technology, we hope to inspire further innovation in the field of
smart home security.

Problem
Formulation
1. Security: One of the primary reasons for implementing facial recognition
technology is to enhance security measures. Facial recognition can be a highly secure
way to authenticate and grant access to a restricted area or building.
2. Convenience: Facial recognition can offer a convenient alternative to traditional
key or card-based entry systems. Users can simply walk up to the door and be
granted access without the need for any physical keys or access cards.
3. Accuracy: Facial recognition technology has become highly accurate in recent
years, making it a reliable means of authentication. This can help prevent
unauthorized access and improve security.
5. Cost-effectiveness: Facial recognition technology can be a cost-effective solution
in the long run, as it eliminates the need for physical keys and access cards.
Additionally, it can reduce the need for security personnel to monitor access points.
6.User-friendly: Facial recognition technology is user-friendly, as users don't need to
carry anything extra to gain access. This can improve the overall user experience and
reduce frustration associated with lost or forgotten access cards.
Overall, a facial recognition door unlock project can offer a highly secure,
convenient, and cost-effective solution for access control, making it a compelling
option for a wide range of settings.

Objectivesof the Work
02
Ensure that the system is secure
and cannot be easily bypassed
by unauthorized users, through
measures such as encryption,
user authentication, and anti-
spoofing techniques.
03
Test the system thoroughly to
ensure that it works reliably in
a range of lighting conditions,
with different camera angles
and distances, and with a
variety of facial features and
expressionsud technology.
01
Develop a facial recognition
system that can accurately
identify authorized individuals
and grant them access to a
locked door.
04
Evaluate the system's accuracy and
speed of recognition, and optimize its
performance through machine learning
and other techniques.
05
Document the project thoroughly, including its
design, implementation, testing, and evaluation, so
that others can replicate or build upon the work in
the future.

Methodology used
The methodologies in this project will include hardware and
software development and integration of both. The
hardware development will involve selecting and integrating
a low-cost camera, microcontroller, and locking
mechanism. In this algorithm will take a few images from
different angles and then check if the user is registered. The
software development will involve developing facial
recognition algorithms, image processing, and database
management. The integration will involve testing and
optimizing our system for accuracy, speed, and reliability

The training of a face-unlocking CNN involves
feeding it thousands of images of faces from a
variety of angles, lighting conditions, and
expressions. One such system is the Face Net
system developed by Google, which uses a deep
convolutional neural network (CNN) to extract
facial features and then applies a distance metric to
determine the similarity between two faces.
Another system is the Deep Face system developed
by Facebook, which uses a deep neural network to
extract facial features and then applies a SoftMax
classifier to classify the faces. Therefore, there is a
need for an alternative algorithm that can provide
high accuracy with low computational costs.
Use of CNN

Convolutional layer:
Activation layer: Pooling layer:
Fully Connected
Layers:
Batch Normalization
layer:
CNN L
CN
ayers

Facial Recognition Door unlock using Machine Learning.

According to the evaluation, the facial recognition system
achieved an overall
accuracy of 95%. This indicates that the system was able to
recognize the
person correctly in most cases.
The FAR was found to be 2%, indicating low susceptibility to
fraudsters. The
FRR is 3%, indicating that the system correctly identified the
matching faces
with high accuracy.
Determines whether it is an authorized person or not based on
the pre-registered
for pre-stored images in the database
Results

Basedon the results of the tests that havebeen carried out, it can
be concluded that the system runs as expected and can be used
as a facial recognition-based automatic door lock system. When
the facial image cannot be recognized or is unsuccessful in being
identified, the door will remain closed. Therefore, the system can
beinstalled asasecurityfeature on adoor to aprivate room.
The latest development of 3D sensors points to a new approach
to facial recognition that may be able to get through the primary
obstacles, such as changes in physical appearance, the ageing
process, position, changes in light intensity, and more generally,
facial emotions, missing data, cosmetics, and occlusions. In the
event of problematic acquisition settings, the geometry
information offered by facial datasets could greatly increase the
accuracy of facial identification.
Results

Based on the results of the tests that have been carried
out, it can be concluded that the system runs as expected
and can be used as a facial recognition-based automatic
door lock system. When the facial image cannot be
recognized or is unsuccessful in being identified, the door
will remain closed. Therefore, the system can be installed
as a security feature on a door to a private room.
The latest development of 3D sensors points to a new
approach to facial recognition that may be able to get
through the primary obstacles, such as changes in
physical appearance, the ageing process, position,
changes in light intensity, and more generally, facial
emotions, missing data, cosmetics, and occlusions. In the
event of problematic acquisition settings, the geometry
information offered by facial datasets could greatly
increase the accuracy of facial identification.
Conclusion
Highlights
1. Facial recognition technology can
provide a convenient and secure
way to unlock doors, eliminating
the need for keys or access codes.
2. Using a camera to capture an
image of the user's face, the
software can compare it to a
database of authorized users to
determine if access should be
granted.
3. This technology can be especially
useful in high-security
environments such as government
buildings, research facilities, or
financial institutions.
4. By integrating facial recognition
with other security systems such as
alarms and cameras, the door
unlock project can provide a
comprehensive security solution

Conclusion
Pitfalls
Facial recognition technology has come under criticism
for potential biases, inaccuracies, and privacy
concerns. It is essential to consider these issues when
implementing the technology and ensure that the
system is fair and transparent.
The accuracy of facial recognition technology can be
affected by various factors such as lighting, angles,
and facial expressions. The system may fail to
recognize authorized users, resulting in frustration and
inconvenience.
It is crucial to implement proper security protocols to
prevent unauthorized access to the facial recognition
system itself, as this can compromise the entire
security system.
Finally, the cost of implementing a facial recognition
door unlock project can be significant, requiring
hardware, software, and ongoing maintenance
expenses. A cost-benefit analysis should be conducted
to ensure that the benefits of the project justify the
investment.

Improved accuracy: As facial recognition technology continues to advance,
there will likely be further improvements in accuracy, particularly in challenging
lighting conditions or with subjects wearing masks or other face coverings.
Scalability: Currently, most facial recognition systems are designed to work with
a relatively small number of users. In the future, it may be possible to scale up
the system to work with larger groups of people, potentially making it useful for
access control in public spaces.
Multi-factor authentication: While facial recognition can be an effective way to
authenticate users, it may be even more secure to combine it with other forms
of authentication, such as passwords or biometric data from other parts of the
body (such as fingerprints or iris scans).
Privacy concerns: Facial recognition technology has come under scrutiny due to
concerns about privacy and potential misuse. Future developments in this area
may focus on ways to mitigate these concerns, such as incorporating encryption
or anonymization techniques.
Future
Scope

References
[1]. O. Deniz, M. Castrillon, M. Hernandez, “Face recognition using independent component analysis and support vector machines,” Pattern Recognition
Letters, vol. 24, pp. 2153-2157, 2003.
[2]. I.J. Cox, J. Ghosn, and P.N. Yianios, “Feature-Based face recognition using mixture-distance,” Computer Vision and Pattern Recognition, 1996.
[3]. M. Turk and A. Pentland, “Eigenfaces for recognition,” J. Cognitive Neuroscience, vol. 3, pp. 71-86, 1991.
[4]. Ho-Man Tang, Michael Lyu, and Irwin King, "Face recognition committee machine," In Proceedings of IEEE International Conference on Acoustics,
Speech, and Signal Processing (ICASSP 2003), pp. 837-
840, April 6-10, 2003
[5]. "Design and Implementation of a Facial Recognition Door Lock System" by D. D. Yang, D. Y. Lee, and J. H. Kim -Link: https://www.mdpi.com/1424-
8220/19/13/3009
[6]."Smart Door Lock using Facial Recognition System" by S. Senthil Kumar, S. Sathish Kumar, and S. Sivakumar.
(https://ieeexplore.ieee.org/document/9041969)
[7]. "Facial Recognition Door Lock Security System using IoT" by M. B. Khan, M. A. Al-Garadi, and A. R. Khan. (https://www.mdpi.com/2504-
3900/25/1/29)
[8]. "Facial Recognition-Based Door Lock System Using Deep Learning" by H. M. Elhoseny, M. A. M. Mahmoud, and S. M.
Riad. (https://www.mdpi.com/1424-8220/20/23/6765)

Facial Recognition Door unlock using Machine Learning.

More Related Content

Similar to Facial Recognition Door unlock using Machine Learning. (20)

Recently uploaded (20)

Facial Recognition Door unlock using Machine Learning.