IoT BASED FACIAL RECOGNITION
SECURITY SYSTEM
Prashanth Balraj Balla
M.E. [Second Year] - Electronics and
Telecommunication Engineering, Alamuri Ratnamala
Institute of Engineering and Technology (ARIET),
Thane - Maharashtra, India
prashanthballa1993@gmail.com
Abstract— In recent years, with the demand for better
security, computers have played a large role. Due to
their precision, large memory banks and high
computing power, considerable development has been
made in the area of face recognition. Computers now
surpass humans in many face recognition tasks. A
human being can remember limited number of faces.
But a computer doesn’t have any limits, and can hence
be used where large databases of facial records are
needed. Such a facial recognition system has many
potential applications including crowd and airport
surveillance, private security and improved human-
computer interaction. Such a system is perfectly suited
to fix security issues and offer flexibility to smart house
control.
This project is aimed to be a complete system for face
recognition: easy to build, cheap cost and effective.
Main purpose is to be set as an alert for home visitors
and provide information about the visitors in a dynamic
website and phone application. It can also be used in
other fields like industries, offices and even air-ports for
identifying wanted people. Among the other bio-metric
techniques, face recognition method offers one great
advantage which is user friendliness.
Keywords— Internet of things (IoT) ,principal
component analysis (PCA) , raspberry pi (RPI) ,global
system for mobile communication (GSM) , local area
network (LAN) ,universal serial bus (USB).
I. INTRODUCTION
Over the last ten years or so, Biometric has
become a popular area of research in computer vision
and one of the most successful applications of image
analysis and understanding. Because of the nature of
the problem, not only computer science researchers
are interested in it, but neuroscientists and
psychologists also.
There are many different traits that can be used as
biometric, including fingerprints, face, iris, retina,
gait, and voice. In general, biometrics are divided
into two broad categories-
Authorized licensed use limited to: VIT University. Downloaded on November 06,2022 at 16:10:30 UTC from IEEE Xplore. Restrictions apply.
Prof.K.T.Jadhao
Assistant Professor,
Alamuri Ratnamala Institute of Engineering and
Technology (ARIET) ,
Thane, – Maharashtra, India
ktjadhao@gmail.com
Physiological biometrics - Those that identify an
individual from an intrinsic physiological or
biological traits (ex. Face, iris, fingerprint, etc.)
Behavioral biometrics – Those that identify an
individual from a behavioral trait (ex. Gait, typing,
etc.)
II. PROBLEM DEFINITION
As in our everyday lives we have information about
everything at our fingertips due to advancement of
various technologies. However we need to know
about the various visitors visiting places ( like office,
home, air-ports, bank-lockers) .We need to track
information about visitors for various scenarios like
1. Entry of unauthorized employees at office
work location.
2. Entry of workers at home.
3. At the place of bank lockers.
4. Violating the security norms at Airport,
Malls.
5. Thief identification at home, Jewelers shop,
Banks, Warehouse.
6. Research laboratories etc.
III. PROPOSED SYSTEM
Using image processing an intelligent security
system can be created which can be used for face
identification. The main objective of this system is to
implement for a particular face and distinguishing it
from a number of stored faces with some real time
variations as well [1].
This system works as explained below-
Whenever a visitor presses the doorbell , an
integrated camera captures picture of it. On this
picture it takes only facial part of it which is required
for feature extraction, which is then further processed
and sends to the database. It tries to match the taken
picture with the stored database in it. Once it is
matched, it gives notification to the owner using IOT
for providing the access. If the image is not matched
then it is stored in Database under unknown visitor
folder for future reference.
When the captured image is matched from the given
database, it identifies the person and send image of
that visitor to the owner of the house via mail. Owner
will get the notification regarding visitor, visitors
image and time of arrival. Now owner decides
whether to allow visitor getting into home or not.
This can be done by sending mail / message to the
security system. Once system receives message from
the owner , it opens the door automatically.
This intelligent system can be further modified to
identify wanted criminals by storing their images in
the database. As soon as it detects their image,
directly it sends message to nearest police station
without the owner’s permission.

IV. ALGORITHM OF PROPOSED SYSTEM

Step 1 : Image will be captured.

Step 2 : Detection of face is done from captured
image.
Step 3: Getting required parameters from image.
Step 4 : Face recognition takes place.
Step 5 : Name of visitor is sent to the owner via Web
, Android notification etc.

Fig1. Block Diagram

V. BLOCK DIAGRAM
In block diagram, it shows various components that
are being used at source. In this, Raspberry Pi model
acts as a centre of the system which will control the VII. PERFORMANCE E VALUATION
action /performance of other devices. It has in built
memory, USB ports, LAN ports. It works at a voltage
of 3.3 volt. Another important thing is camera of Training process: We have trained some images of
5MP, which is going to capture the image of visitor known persons in database. At the time of training it
and it is connected to the RPI. Algorithm for this, is collects maximum number of images of the person.
proposed to be implemented on python which is Larger the number of samples, more will be the
again in built software of RPI accuracy of system.
RPI sends information about visitor to the owner
using IOT i.e., via web notification or android Recognition process: As soon as visitor presses
notification. After receiving reply from owner, the doorbell, camera captures image and tries to match
doors will opened / or remain closed. with the stored database in it. If match found then it
sends mail like as shown in fig. with visitors name as
VI. FLOWCHART OF PROPOSED SYSTEM well as image of visitor.
When someone presses the doorbell, camera captures
picture of that person. From that images, it extracts Granting permission: After receiving mail of visitor
the desired parameters and checks with available information, owner decides whether to grant access
database. or not by sending message as shown in fig. If he
If match found, then it sends message to owner of the sends yes then doors will be opened and if sends no
house about visitors name, visiting timing etc and it then doors will remain closed only.
takes permission from owner to permit visitor or not.
If match did not found, then it saves that image of
new visitor to the new folder, so that it may be
helpful in recognizing that visitor in future [2].

Authorized licensed use limited to: VIT University. Downloaded on November 06,2022 at 16:10:30 UTC from IEEE Xplore. Restrictions apply.
 Fig.3.Training of image

Fig.4.mail sent to owner

Authorized licensed use limited to: VIT University. Downloaded on November 06,2022 at 16:10:30 UTC from IEEE Xplore. Restrictions apply.
 VIII. CONCLUSION
Internet of things is the need of the era. Proposed
system integrates the IoT as a backbone and
biometric face recognition as one of the tool for
security purpose [3]. A biometric technique which
can be used for identification and verification of
person is proposed. The goal of our project is to
benefit and protect human lives as well as to get the
information about visitors of desired location to the
owner using IoT.
The system can be used at various locations like
residences, offices, industries, bank lockers, airports
for various purposes like theft identification,
surveillance etc.

ACKNOWLEDGMENT
It is a great pleasure and a moment of immense
satisfaction for me to express my profound gratitude
to my project guide Prof. K.T.Jadhao, Assistant
Professor, Electronics and Telecommunication
Engineering Department whose constant
encouragement enabled me to work enthusiastically.
His perpetual motivation, patience and expertise in
discussion during progress of work have benefited
me to an extent, which is beyond expression.
Working under his guidance has been a fruitful and
unforgettable experience. Despite of his busy
schedule, he was always available to give me advise,
Fig.5.sending message support and guidance during entire period of my
project and publishing this paper. The completion of
this project and paper would not have been possible
without his constant support and patience guidance.

REFERENCES
[1] Ayman Ben Thabet , Nidhal Ben Amor - Enhanced smart
doorbell system based on face recognition
[2] Liton Chandra Paul and Abdulla Al sumam – Face
Recognition Using Principal Component Analysis Method:
[3] - Jump up^ Vermesan, Ovidiu; Friess, Peter (2013). Internet
of Things: Converging Technologies for Smart Environments
and Integrated Ecosystems (PDF). Aalborg, Denmark: River
Publishers. ISBN 978-87-92982-96-4.
[4] DWT-based face recognition using fast walsh hadamard
transform and chiral image superimposition as pre-processing
techniques- G V Niveditha, B P Sharmila, K Manikantan
[5] Face Recognition Using Active Illumination Equalization
and Mirror Image Superposition as Pre-processing Techniques -
S.Hitesh Babu ; H.R.Shreyas ; K.Manikantan
[6] - Real time face recognition engine using compact features
for electronics key- Gede Pasek Suta Wijaya ; Ario Yudo
Husodo ; Andy Hidayat Jatmika
[7] - Brown, Eric (13 September 2016). "Who Needs the
Internet of Things?". Linux.com. Retrieved 23 October 2016.
[8] - Jump up^ "Internet of Things: Science Fiction or Business
Fact?" (PDF). Harvard Business Review. November 2014.
Retrieved 23 October 2016.
[9] - Jump up^ "An Introduction to the Internet of Things
(IoT)" (PDF). Cisco.com. San Francisco, California: Lopez
Research. November 2013. Retrieved 23 October 2016.
Fig 6. output of granting access

Authorized licensed use limited to: VIT University. Downloaded on November 06,2022 at 16:10:30 UTC from IEEE Xplore. Restrictions apply.