IEEE - 56998

Development of Face Recognition-Based Smart

2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT) | 979-8-3503-3509-5/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICCCNT56998.2023.10307437

Door Lock System with Remote Servo Control

Authentication
Chaitanya Kolluru Akhil G V Siva Priyanka S
Department of ECE Department of ECE Department of ECE
CBIT, Hyderabad CBIT, Hyderabad CBIT, Hyderabad
India - 500090 India - 500090 Hyderabad, India - 500090
[email protected] [email protected] sivapriyankas [email protected]

Krishna Reddy D Aruru Sai Kumar

Department of ECE Department of ECE
CBIT, Hyderabad VNR VJIET, Bachupally
India - 500090 Hyderabad, India - 500090
dkreddy [email protected] [email protected]

Abstract—Security in the home is more crucial than ever for of devices that work together to process and share information,
safeguarding personal belongings and avoiding identity theft. increasing efficiency and intelligence [5]-[8].
Key loss or theft occurrences have been shown to be vulnerable
to traditional access control measures like keys, security cards,
passwords, or patterns, which can result in major security Facial recognition is one of the most popular biometric
breaches. To solve this problem for more effective and safe methods because of its stability and high level of security.
door access control systems, a novel method based on face Faces contain multiple features, making them more difficult to
recognition employing machine learning and IoT technologies forge, steal, or borrow. Face recognition systems work by first
have been presented. A camera records photographs of people detecting faces in images captured by cameras. The person can
trying to enter the building, and the Raspberry Pi, a small
programmable computer board, acts as the main controller for be recognized without physical contact with any specialized
the face recognition, youth system, and locking system. IoT hardware. Machine learning techniques are commonly used
technology enables users to manage door access and keep an to improve the accuracy of facial recognition systems over
eye on activity from a distance. This novel approach boosts time. With deep learning, the system is trained using images
protection, lowers the danger of theft, and provides a practical, of authorized users, and the accuracy of recognition improves
adaptable, and scalable solution for a variety of home security
applications. automatically [9]-[12].
Index Terms—Face Recognition, Raspberry PI, OpenCV, PIR
sensor, Internet of Things, surveillance. Home security is crucial for ensuring the safety of loved
ones and property. Traditional security methods such as keys
I. I NTRODUCTION and passwords have limitations, leading to the adoption of
In today’s world, ensuring the safety and protection of loved biometric systems, such as face recognition. This method is
ones and property is crucial, and home security systems are highly secure and uses deep learning techniques to improve
essential for achieving this. While traditional methods like accuracy. IoT technology can be used to control door access
keys, passwords, and ID cards were used in the past, they and send notifications. The Blynk app is used to design a
had limitations, and biometric technology has emerged as a graphical interface for controlling door access and sending
promising alternative. Biometric systems [1]-[4] use unique notifications to smart devices. The program is very helpful in
and measurable individual characteristics for identification, Internet of Things (IoT) applications because it can be linked
such as fingerprints, palm prints, and DNA analysis, making to the internet using Wi-Fi, Ethernet, or an ESP8266 chip.
them more secure than traditional methods. By communicating The software allows users to hit the doorbell in the event that
with each other, these objects can form interconnected systems the face recognition system is unable to identify them, which
that can be remotely controlled and monitored. The result results in a notification being delivered to the homeowner’s
is a more efficient and convenient way to manage various smartphone[13]-[15]. A live video stream is included in the
aspects of daily life, from home automation to industrial message to help identify the individual attempting to open the
control systems. By incorporating the Internet of Things (IoT) door. A Raspberry PI 4 and a PICAMERA are utilized to
technology, smart systems can create a large or small network capture images upon motion detection by a PIR sensor.

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi
 IEEE - 56998

II. RELATED WORKS and turn on an LED light so that an image can be taken even
IoT devices play a crucial role in IoT-based systems. While at night[23]-[25]. Only if motion is detected during the night
previous studies have explored various methods for home will the system proceed to facial detection. The method for
security, the use of Gmail as a medium for conveying security face detection will process the current frame with the detected
information is still relatively underutilized. Researchers have motion if motion is found. The flowchart for the proposed
shown interest in utilizing Gmail to monitor system activities. system is shown below.
The fact that this software is free makes it simple for the gen-
eral public to develop it. Android-powered smartphones can
monitor home security systems and use PIR sensors to identify
human presence. As notifications are sent through the Gmail
app, these sensors listen for passive infrared rays released by
the human body [16]. Blynk is a popular IoT platform that
allows users to create and manage IoT applications with a
simple drag-and-drop interface. One of the features of Blynk
is the ability to send notifications to users through various
channels, including email, SMS, and push notifications[17]. In
the context of the text provided, Blynk could be used to create
an IoT application that collects data from wireless sensors in
a network, such as temperature or humidity sensors in a smart
home, and sends notifications to the user’s Gmail account.
The notifications through Gmail can provide you with
important information about the state of your IoT devices,
allowing you to take timely action. For instance, if you receive
an email notification from Blynk that the humidity in your
basement has reached a high level, you can take steps to
address the issue before it leads to damage or mold growth[18].
Similarly, if you receive an email notification about a security
sensor being triggered, you can take appropriate action, such as
checking the security camera footage or contacting authorities
if necessary. By using Blynk to notify you through Gmail,
you can stay connected to your IoT devices and receive
timely updates, allowing you to monitor and manage your Fig. 1. Workflow of proposed Smart Door Lock System
IoT applications effectively[19]. This can help you optimize
the performance of your IoT network, ensure the safety and
security of your environment, and enhance your overall IoT The Raspberry Pi 4B, PIR Sensor, Raspberry Pi Camera,
experience[20]. Smartphone, Pen Drive, and Power Supply are among the
Face recognition technology has been used in a number of hardware elements employed in the study project. The GIPO
IoT applications, including unmanned aerial vehicles, smart pins on the Raspberry Pi, a diminutive controller with USB
classrooms, home security systems, smart houses, and smart modules, are used to connect the PIR motion sensor and
surveillance. Conventional techniques including local binary Raspberry Pi Camera. The Raspberry Pi is also connected to
patterns, neural networks, support vector machines, and k- the smartphone through the internet. The Raspberry Pi serves
nearest neighbours have been employed in prior IoT facial as the central processing unit for capturing and processing
recognition implementations [21]. The paper is presented as images, as well as sending notifications[26]-[30]. The images
the first proposed security door lock system followed by face are captured using the Pi-Camera module, and the PIR motion
detection and recognition using am and results and discussions sensor detects motion.
and finally concluded the discussions[22]. IV. FACE DETECTION & RECOGNITION USING
III. PROPOSED SECURITY DOOR LOCK SYSTEM MACHINE LEARNING
The proposed system entails taking photos with a camera A technique based on cascade object detection is used to
depicted in Fig. 1. A computer vision module will be used to detect a face in an acquired image, as shown in Fig 2. This
detect faces in the photos after they have been taken. Images detector recognizes the image and extracts the facial region
will be emailed to a smartphone as soon as faces are found from it. In a separate environment, the facial recognition
in them. This technique is more effective at defending an area system keeps photos of an authorized person’s face. Nor-
while being less cost- effective. Face detection and motion malizing the facial photographs eliminates undesired details
detection are the two categories into which the application can produced by lighting limits during acquisition, increasing their
be split. The PIR sensor’s primary function is to detect motion quality. The facial recognition method will be carried out

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi
 IEEE - 56998

with the detected face and image saved in the database. values in dark regions is subtracted from the sum of pixel
Face recognition in OpenCV can be accomplished using a values in light regions to compute these characteristics and is
variety of algorithms, including feature-based and model- defined as in Eq. (1)
based algorithms. In real-time security systems, feature- based
algorithms are commonly used. Principal Component Analysis F V = P D–P L (1)
(PCA) and Linear Discriminant Analysis (LDA) are useful
Where, FV = Feature Value, PD = Sum of pixels in the dark
methods for extracting features for recognition[31]-[35]. Both
region, PL = Sum of pixels in the light region.
strategies reduce the feature base procedure in the same way,
The resulting feature value is compared to a threshold to
however, when large training sets are employed in recognition,
determine if a face candidate is present in the sub-window.
Linear Discriminant Analysis (LDA) surpasses Principal Com-
The third phase of the algorithm is AdaBoost training [12].
ponent Analysis (PCA). LDA distinguishes the majority of the
AdaBoost is a machine learning algorithm that is used to select
information in photos by computing intra-class and interclass
a subset of the most relevant Haar-like features from a large
dispersion matrices. It runs the LDA and distinguishes between
pool of features.
normalized photos stored in the database and identified images
Finally, the cascade classifier is used to scan the entire im-
based on the database-stored images. LDA validates the image
age with the Haar-like features a window in various scalesand
and gives extra signals to the system, allowing the person to
orientations to detect face candidates. When a face candidate
be identified. The person can be correctly identified using
is detected, the captured image is processed further to verify if
OpenCV and Anaconda software- based image processing
it meets the criteria of a human face, such as the presence of
techniques [36].
head and shoulder features. If the criteria are met, the captured
image is sent as a notification to the user’s Smartphone as a
face detection event. Otherwise, the image is discarded as a
non- face. This process ensures that only valid face detections
are sent as notifications to the user’s Smartphone[41]-[44].
B. Dlib’s face recognition
Dlib’s face recognition algorithm is based on deep metric
learning using a neural network architecture called ResNet.
This approach learns a mapping from face images to a high-
dimensional feature space, where faces from the same per-
sonare closer to each other than faces from different people.
In the process of face recognition, the key challenge is to
identify unique facial features. One approach is to compare the
unknown face with all faces in a database, which can be slow
and impractical for large datasets[45]-[46]. Cloud storage can
help, but it can lead to delays in processing time. For projects
that require quick recognition, this approach is not suitable
due to its inefficiency and high computational complexity.
Dlib’s face recognition algorithm works in several steps:
1) Face detection: The first step is to detect faces in an
input image. Dlib uses a variant of the Histogram of Oriented
Gradients (HOG) method to detect faces. Given an input image
I, the face detection step can be represented as a function f(I)
that returns the location of the faces in the image and is defined
in Eq. (2)
Fig. 2. Face Detection Recognition Flow Chart
f (I) = [(x1 , y1 , w1 , h1 ), (x2 , y2 , w2 , h2 ), ...] (2)
where (xi , yi ) is the top-left corner of the ith face bounding
A. Haar Cascade box, wi is the width of the bounding box, and hi is the height
Face detection is a useful function in a variety of ap- of the bounding box.
plications. Haar cascade feature detection is a popular face 2) Face alignment: Once a face is detected, the algorithm
detection approach that can recognize single or multiple faces. aligns it by finding the facial landmarks such as the corners of
In the suggested system, the OpenCV library, a widely used the eyes, nose, and mouth, as shown in Fig 4. Given the face
open-source library for image processing, is employed for face bounding boxes from step 1, the face alignment step can be
detection[37]-[40]. The face detection method uses Haar-like represented as a function g(I, [(xi , yi , wi , hi )]), which returns
properties to recognize frontal human faces. The sum of pixel the aligned faces. This can be expressed as:

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi
 IEEE - 56998

The Pi Camera successfully captures photographs, verifies

g(I, [(xi , yi , wi , hi )]) = [F1 , F2 , ...] (3) them against the given dataset, and performs face identification
on these images. The Dlib algorithm, specifically the Haar
where F1 , F2 in Eq. (3) is the aligned face corresponding cascade face detection technique, is used for face recognition
to the ith bounding box. and is notable for its real-time performance and minimal CPU
3) Face representation: The algorithm then extracts a set of resource utilization. The system works in real-time and has
128-dimensional feature vectors from the aligned face image, four major functions: motion detection, image capture, human
known as embeddings. Given the aligned faces from step 2, face detection, and result/notification delivery to the user’s
the face embedding step can be represented as a function smartphone. The camera records events and photographs, and
h([F1 , F2 , ...]), which returns the face embeddings. This can notifications are delivered to a smartphone app, in this example
be expressed as: the ”Gmail” app, as illustrated in Fig. 4.
h([F1 , F2 , ...]) = [e1 , e2 , ..., ei ] (4)
where ei in Eq. (4) is the face embedding corresponding to
the ith aligned face.
4) Face comparison: The algorithm computes the Eu-
clidean distance between their feature vectors to compare two
faces. The faces are considered a match if the distance is below
a certain threshold.
Given two face embeddings e1 and e2 , the face comparison
step can be represented using a distance metric such as
Euclidean distance [15] or cosine similarity is represented in
Eq. (5). This can be expressed as:
di(e1 , e2 ) = ||e1 − e2 || (5)
If the distance between two face embeddings is below a
certain threshold, they are considered to be the same person.
Fig. 4. Intruder entry will be notified Via Gmail.
Otherwise, they are considered to be different people.
5) Face recognition: To recognize a person, the algorithm
compares an input face to a set of reference faces, each The Raspberry Pi 4, used in the system, has built-in Wi-Fi
represented by their feature vector. and Bluetooth capabilities, allowing for remote monitoring of
motion events and instant display of images on the smartphone
V. R ESULT A NALYSIS device. Blynk is an IoT platform that allows users to remotely
Blynk, a web application [15-16], was used to improve manage Microcontrollers, MicroProcessors, and NodeMCU
the functionality of the proposed smart face-detecting security through the internet via iOS and Android smartphones. Users
system for homes and offices. The experimental setup, as illus- can choose to use the Blynk Cloud or run their own private
trated in Fig.3, contains the essential hardware configuration Blynk server. When a user sends a command from the Blynk
as well as software installation. app, it is processed by the Blynk server, and data is delivered
to the hardware.
Fig 5, Fig. 6 are the images of known and unknown persons
detecting the proposed work. Fig. 7 depicts the graph between
Acuuracy and Runtime Plot. The accuracy attained for various
picture counts used to train the model is shown in TABLE I.
The precision attained in poor light is described in TABLE II.
Based on the separation between the camera and the subject,
the accuracy of facial recognition is described in TABLE III.
VI. C ONCLUSION
In conclusion, this research work is to develop a smart
security system using face recognition with remote door lock
control. The system uses a Raspberry Pi board with a camera
module, PIR sensor, servo motor, and Blynk application for
remote control. The implementation involves the use of Haar
Cascade and Dlib’s face recognition algorithms for face de-
Fig. 3. Proposed Model Connections
tection and recognition, a PIR sensor for motion detection in
low-light conditions, and a servo motor for door lock control.

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi
 IEEE - 56998

TABLE I
ACCURACY RATE.

S.No Image count is given Accuracy

for training the model
1 50 75.74%
2 100 84.35%
3 200 89.56%
4 500 91.9%
5 1000 92.72%

TABLE II
ACCURACY RATE IN LOW LIGHT

S.No Image count is given Accuracy

for training the model
1 100 69.7%
Fig. 5. Images of known and unknown are recognized 2 500 75,61%
3 700 80.24%

TABLE III
DISTANCE-BASED ACCURACY

S.No Distance (in cm) Accuracy Status

1 30 89.8% Device Unlocked
2 60 84.94% Device Unlocked
3 85 83.7% Device Unlocked
4 100 79.53% Device Locked
5 130 74.62% Device Locked

user in case an intruder is detected by the system. This is

achieved by sending an email notification to the user’s email
address. The system was evaluated for accuracy, performance,
and reliability through various metrics such as face recognition
Fig. 6. Face recognition of the known person. accuracy, response time, and notification reliability. On 1000
images of the dataset, an accuracy of 92.72% is achieved.
The average accuracy is about 88% to 90%. The application
work over a range of 30 to 100 cm. During low light, it
acquired an accuracy of 80.24%. The system can be made
better by adding liveliness to the facial recognition module,
which only recognizes real faces and leaves phone photos of
verified individuals.

R EFERENCES
[1] S. Siva Priyanka and T. Kishore Kumar, “Signed Convex Combination
of Fast Convergence Algorithm to Generalized Sidelobe Canceller Beam-
former for Multi-Channel Speech Enhancement,” Traitement du Signal,
vol. 38, no. 3, pp. 785-795, June 2021.
[2] S. S. Priyanka, “A review on adaptive beamforming techniques for speech
enhancement,” 2017 Innovations in Power and Advanced Computing
Technologies (i-PACT), India, pp. 1-6,2017.
[3] Priyanka, S. Siva, and T. Kishore Kumar. “Generalized Sidelobe Canceller
Beamforming with Combined Postfilter and Sparse NMF for Speech
Enhancement”, Fluctuation and Noise Letters, Vol. 20, no. 02, 2021.
Fig. 7. Accuracy vs Runtime plot [4] Priyanka, S.S., Kumar, T.K. “Multi-channel speech enhancement us-
ing early and late fusion convolutional neural networks,” SIViP 17,
973–979,2023.
[5] S. S. Priyanka and T. K. Kumar, “Adaptive Beamforming Using Zelinski-
TSNR Multichannel Postfilter for Speech Enhancement,,” 2018 9th
Email notification and remote servo control are the main International Conference on Computing, Communication and Networking
novelties. The email notification feature is used to alert the Technologies (ICCCNT), Bengaluru, India, pp. 1-6, 2018.

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi
 IEEE - 56998

[6] S. S. Priyanka and T. K. Kumar, “GSC Beamforming using Different [26] Aruru Sai Kumar, T.V.K. Hanumantha Rao, “Scalable benchmark syn-
Adaptive Algorithms for Speech Enhancement,,” 2019 10th International thesis for performance evaluation of NoC core mapping”, Microproces-
Conference on Computing, Communication and Networking Technolo- sors and Microsystems, Volume 79, November 2020.
gies (ICCCNT), India, pp. 1-6, 2019. [27] Aruru Sai Kumar, T.V.K. Hanumantha Rao, “An Adaptive Core Mapping
[7] Xingsi Xue et al., “Design and Performance Analysis of 32 × 32 Memory Algorithm on NoC for Future Heterogeneous System-on-Chip”, Comput-
Array SRAM for Low-Power Applications”, Electronics, Vol.12, Issue. ers and Electrical Engineering, Volume 95, October 2021.
4, pp.834, 2023. [28] A. S. Kumar, T. V. K. Hanumantha Rao and B. N. Kumar Reddy, “Exact
[8] S. R. Pillutla et al., “Mongoose OS Prototyped ESP8266 Based Weather Formulas for Fault Aware Core Mapping on NoC Reliability,” 2020 IEEE
Forecaster,” 2022 International Conference on Recent Trends in Mi- 17th India Council International Conference (INDICON), pp. 1-5, 2020.
croelectronics, Automation, Computing and Communications Systems [29] A. S. Kumar et al., “Performance and communication energy constrained
(ICMACC), Hyderabad, India, pp. 371-375, 2022. embedded benchmark for fault tolerant core mapping onto NoC architec-
[9] V. B. Sreenivasulu, V. Narendar, “Junctionless SOI FinFET with advanced tures” , International Journal of Ad Hoc and Ubiquitous Computing, vol.
spacer techniques for sub-3 nm technology nodes”, AEU - International 41, no. 2, pp. 108-117, 2022.
Journal of Electronics and Communications, vol. 145, pp. 154069, 2022. [30] A. S. Kumar and T. V. K. H. Rao, “Efficient Core Mapping on
[10] V. B. Sreenivasulu et al,, “Common Source Amplifier and Ring Os- Customization of NoC Platforms,” 2019 IEEE International Symposium
cillator Performance Optimization Using Multi-Bridge/Nanosheet FETs”, on Smart Electronic Systems (iSES) (Formerly iNiS), pp. 57-62, 2019.
ECS Journal of Solid State Science and Technology, 2023. [31] Kumar, A.S., Naresh Kumar Reddy, B. “An Efficient Real-Time Em-
[11] V. B. Sreenivasulu et al.,“Design Insights of Nanosheet FET and CMOS bedded Application Mapping for NoC Based Multiprocessor System on
Circuit Applications at 5-nm Technology Node,” IEEE Transactions on Chip” , Wireless Personal Communications, vol. 128, pp.2937–2952,
Electron Devices, vol. 69, no. 8, pp. 4115-4122, Aug. 2022. 2023.
[12] S. Ramakrishna Pillutla, et al., “An Efficient two Dimensional Map- [32] Aruru Sai Kumar and T. V. K. Hanumantha Rao, “An efficient low
ping of External Surroundings for Robotics and Drones,” 2022 13th latency router architecture for mesh based NoC” , in International
International Conference on Computing Communication and Networking Conference on Communications, Signal Processing and VLSI (IC2SV-
Technologies (ICCCNT), India, pp. 1-6, 2022. 2019), India, pp. 241-248, 2021.
[33] Aruru Sai Kumar, T.V.K. Hanumantha Rao, “Performance assessment
[13] V. P. Brahmaiah et al., “An Efficient Pulse Detector with Arduino
of adaptive core mapping for NoC-based architectures” , International
and IR Sensor,” 2022 13th International Conference on Computing
Journal of Embedded Systems, Vol. 15, No. 5, 2022.
Communication and Networking Technologies (ICCCNT), India, pp. 1-6,
[34] B. N. K. Reddy, A. James and A. S. Kumar, “Fault-Tolerant Core Map-
2022.
ping for NoC based architectures with improved Performance and Energy
[14] S. S. Priyanka, et al., “An Efficient Door Access System Controlled Efficiency,” 2022 29th IEEE International Conference on Electronics,
by Temperature using LabVIEW,” 2022 13th International Conference Circuits and Systems (ICECS), UK, pp. 1-4, 2022.
on Computing Communication and Networking Technologies (ICCCNT), [35] A. S. Kumar et al., “An Efficient AVR interfaced Bluetooth controlled
India, pp. 1-6, 2022. Robotic Car system,” 2023 13th International Conference on Cloud
[15] Raghava Rao, K., Naresh Kumar Reddy, B. & Kumar, A.S. “Using Computing, Data Science & Engineering (Confluence), India, pp. 499-
advanced distributed energy efficient clustering increasing the network 502, 2023.
lifetime in wireless sensor networks” , Soft Computing, 2023. [36] A. S. Kumar et al., “A Novel RRAM-based FPGA architecture with
[16] Wahyuni, Refni, Aditya Rickyta, Uci Rahmalisa, and Yuda Irawan. Improved Performance and Optimization Parameters,” 2022 IEEE 19th
“Home security alarm using Wemos D1 and HC-SR501 sensor-based India Council International Conference (INDICON), Kochi, India, pp. 1-
telegram notification.” Journal of Robotics and Control (JRC) 2, no. 3 5, 2022.
(2021): 200-204. [37] B. N. K. Reddy and A. S. Kumar, “An Efficient Low-Power VIP-based
[17] Rif’an, Muhammad. “Internet of things (IoT): Blynk framework for the VC Router Architecture for Mesh-based NoC,” 2022 IEEE 19th India
smart home.” KnE social sciences (2019):579-586. Council International Conference (INDICON), India, pp. 1-5, 2022.
[18] Balakrishnan, Biju, Punit Suryarao, Rashmi Singh, Sakshi Shetty, and [38] S. R. Pillutla et al., “Mongoose OS Prototyped ESP8266 Based Weather
Sparsha Upadhyay. “Vehicle Anti-theft Face Recognition System, Speed Forecaster,” 2022 International Conference on Recent Trends in Mi-
Control and Obstacle Detection using Raspberry Pi.” In 2022 IEEE 5th croelectronics, Automation, Computing and Communications Systems
International Symposium in Robotics and Manufacturing Automation (ICMACC), Hyderabad, India, pp. 371-375, 2022.
(ROMA), pp. 1-5. IEEE, 2022. [39] Kumar, A.S. and Kumari, B.P., “Implementation of High speed and Low
[19] Khairuddin, M. H., S. Shahbudin, and M. Kassim. “A smart building power 6T SRAM cell using FinFET at 22nm technology”, International
security system with intelligent face detection and recognition.” In lop Journal of Engineering And Computer Science(IJECS), Vol. 5 No. 12,
conference series: Materials science and engineering, vol. 1176, no. 1, p. 2016.
012030. IOP Publishing,2021. [40] S. S. Priyanka and T. K. Kumar, “GSC Adaptive Beamforming Using
[20] Singh, Arpit, Saumya Bhatt, Vishal Nayak, and Manan Shah. “Automa- Fast NLMS Algorithm for Speech Enhancement,” 2019 3rd International
tion of surveillance systems usingdeep leaming and facial recognition.” Conference on Imaging, Signal Processing and Communication (ICISPC),
International Journal of System Assurance Engineering and Management, pp. 160-165,2019.
1- 10, 2023. [41] A Sai Kumar et al., “A Robust Algorithm for Digital Image copyright
[21] Yedulapuram, Sharvani, Rajeshwarrao Arabelli, Kommabatia Mahender, protection”, International Journal of Electrical and Electronic Engineering
and Chintoju Sidhardha.“Automatic Door Lock System by Face Recog- Telecommunications, Vol. 3, No. 3, 2014.
nition.” In JOP Conference Series: Materials Science and Engineering, [42] Rao, N., Rao, T.S. and Kumar, A.S., “Analysis for saturation throughput
vol. 981, no. 3, p. 032036. IOP Publishing, 2020. using various mechanisms for IEEE 802.11 WLAN’s”, Int. J. Res. Appl.
[22] Aydin, Ilhan, and Nashwan Adnan Othman. “A new loT combined face Sci. Eng. Technol, 2(5), 2015.
detection of people by using computer vision for security application.” In [43] Sreenivasulu, V. Bharath, et al.,“Characterization and optimization of
2017 International Artificial Intelligence and Data Processing Symposium junctionless gate-all-around vertically stacked nanowire FETs for sub-5
(IDAP), pp. 1-6. IEEE, 2017. nm technology nodes”, Microelectronics Journal, 116, 2021.
[23] G. Shanthi et al., “An Efficient FPGA Implementation of Cascade [44] Sreenivasulu V and Narendar V., “Design and Deep Insights into Sub-10
Integrator Comb Filter,” 2022 International Conference on Intelligent nm Spacer Engineered Junctionless FinFET for Nanoscale Applications”,
Innovations in Engineering and Technology (ICIIET), pp. 151-156, 2022. ECS J. Solid State Sci. Technology, 10 ,2021.
[24] A. Sai Kumar, U. Siddhesh, N. Sai kiran and K. Bhavitha, “Design of [45] Arulananth, T.S. et al., “Evaluation of low power consumption network
High Speed 8-bit Vedic Multiplier using Brent Kung Adders,” 2022 13th on chip routing architecture”, Microprocessors and Microsystems, Volume
International Conference on Computing Communication and Networking 82, 2021.
Technologies (ICCCNT), pp. 1-5, 2022. [46] G. Shanthi, A. Sai Kumar, U. Bhargava Sai, S. Sajida and G. Karthikeya,
[25] A. Sai Kumar et al., “Characterization for Sub-5nm Technology Nodes “Tunable design of OTA2-C2 Low Pass Filter using CMOS Technology:
of Junctionless Gate-All-Around Nanowire FETs,” 2022 13th Interna- An Innovative Approach,” 2023 4th International Conference for Emerg-
tional Conference on Computing Communication and Networking Tech- ing Technology (INCET), Belgaum, India, pp. 1-6, 2023.
nologies (ICCCNT), pp. 1-5, 2022.

14th ICCCNT IEEE Conference

Authorized licensed use limited to: JNT University Kakinada. Downloaded on January 24,2024 at 09:53:16 UTC from IEEE Xplore. Restrictions apply.
July 6-8, 2023
IIT - Delhi, Delhi