You are on page 1of 4

FACE RECOGNITION AND ATTENDANCE MARKING

USING MACHINE LEARNING

1st S.V.N.Rahul
2nd Parikshit Sharma 3rd Priyanka
Cse.Iot Chandigarh University
Cse.Iot Chandigarh University Cse.Iot Chandigarh University
Mohali Punjab India
Mohali, Punjab India Mohali, Punjab India
19BCS4539@cuchd.in
19BCS4520@cuchd.in 19BCS4514@cuchd.in

Abstract—In the wake of the COVID-19 Pandemic, many aca- demic


institutions, Government organisations and Private sectorsare developed in the UK by Foundation.devices typically consist
suffering from the global cause which immensely affected. This of a microprocessor, memory unit , and some input/output
spreads quickly through close contact with anyone . This project aims ports, all packed into a credit-card-sized form factor. They are
at face recognition and attendance marking which will play a major capable of running a many operating systems and softwares ,
role in attendance system and also acts as a transparent identity proof making them versatile tools for a range of applications. It
in which the proxy is eliminated and this project can be used not only
was initially created as an affordable as well as accessible
in the education industry and many more like Industries,Hospitals etc
way to involve students and hobbyists for a wide range of
With the fast-moving world and everything moving towards
projects. Machine Learning Algorithm: The algorithms are
automation so we have to update and we have to transform the life
style towards automation as the new technologies are booming and computer programs that can improve their performance on a
many things. particular task through experience on there own . They are
used in a wide range of applications, such as natural
. Python programming ,tensor flow ,the open CV library,Keras
Camera module for face detection,Display Module Or Screen are the language processing, computer vision, speech recognition, and
technical and software domains used in fabricating the project ( Face recommendation systems.Here we make use of the computer
Detection and attendance marking). vision advantage of the algorithm as follows. The machine
Index Terms—Python,Tensor Flow ,Open CV, 4,Ml( Machine learning algorithm is utilized to process the images captured
learning) algorithm ,Camera module by the Attendance marking camera. The algorithm is trained
to recognize whether a face recognition is present in the image
I. INTRODUCTION or not and also that face recognition is worn by the person
Face detection and attendance marking using Machine in image properly.
learning is a project that utilizes computer vision technology Display: The Attendance marking supports a variety of
to identify whether a person is present or not. In this project, display modules, including HDMI displays, composite video
camera is used to capture images of people, and then these displays, and displays that connect to the GPIO pins.In our
images are processed using machine learning algorithms to project As the machine learning algorithm has processed the
determine whether or not a face recognition is present. image and determined whether a face recognition is present
The project involves some main components: or not, the result is displayed on a screen or other output
Attendance marking: This is a embedded , low-cost device.
computer that provides the processing power for the project. It II. RELATED WORK
is connected with a camera module that can capture images of
people with or without detection.It is a series of small single- R. K. Kodali et al. [1] proposed that to prevent the spread
board computers of COVID-19, it is necessary to develop a real-time face
recognition detection model using a basic Convolutional
Chandigarh University Neural Network (CNN). This model is built using TensorFlow,
Keras, Scikit-learn, and OpenCV. The proposed work consists
of three stages: pre-processing, training the CNN, and real-
time classification. The CNN classifies faces with and without
detections, with a validation accuracy of 96
E. Soumya et al. [2] proposed that recognizing whether
someone is wearing a face recognition is a critical task
with many applications. Automated face recognition detection
algorithms
currently available are designed to detect whether a person K. P, S. Upadhyay . [?], [?]An automated system has been
is wearing a detection or not. Deep learning has significantly developed to address the limitations of manual temperature
improved face recognition recognition, with three algorithms and detection scanning in public spaces. The system uses con-
being commonly used: You Only Look Once (YOLO), tactless temperature scanning and a camera to detect high
Convolutional Neural Networks (CNN), and Recurrent Neural temperatures or the lack of detections. A gate-like
Networks (RNN). These algorithms can help identify mechanism is employed to block entry if feverishness or no
whether someone is wearing a face recognition in public detection is detected. The system utilizes a temperature sensor
places and provide results. and camera connected to a Attendance marking to automate
J. N. C, et al. [3]Wearing a detection and maintaining a 6- the entire COVID scanning process, reducing the risk of
foot distance between people is crucial to prevent the spread COVID-19 transmission in crowded locations such as malls,
of COVID-19. The need for protective equipment, particularly schools, and universities.
face recognitions, has become essential due to the global D. Kolosov . [10] D.Kolosov proposed that the development
health crisis. This project aims to create a real-time face of efficient on-the-edge deep learning (DL) applications is
detection and identification system using machine learning and challenging due to the need to explore different DL models
a Convo- lutional Neural Network (CNN) model developed with varying trade-offs, optimization options, and edge
with Tensor Flow, Keras, Scikit-learn, and OpenCV. The devices with different constraints. To address this issue, a case
proposed method also includes Principal Component Analysis study was conducted where seven Image Classification (IC)
(PCA) and the HAAR Cascade Algorithm. A green rectangle and six Object Detection (OD) SOTA DL models were
is displayed around a detectioned face, while a red rectangle evaluated on five edge devices. The models were optimized
appears aroundan undetectioned face in the video stream. using SOTA optimization frameworks and different
S,Sruthi et al. [4] Proposed that it is crucial to prevent optimization options, and the devices were compared based on
the spread of the virus, but some people forget to wear inference time, value, and efficiency. The study provides
them, leading to rapid spread of the sickness. To detect valuable insights into selecting the appropriate optimization
detection usage in public areas, a face recognition detector is framework, library, and device depending on the target metric.
necessary. The WHO and governments have implemented M. S. Patil . [11] This project proposes a contactless face
measures such as detection-wearing, sanitation, vaccination, recognition detection model based on machine learning and
and social distancing toprevent the spread of COVID-19. image processing technology to accurately identify if a person
M. A. L. Pagarigan . [5] stated that research project has is wearing a detection properly or not. It follows social
developed a lightweight CNN model that uses a camera to distancing protocols and can be installed in public places like
detect whether people are wearing face recognitions correctly ATMs, malls, educational institutions, and offices to reduce
or incorrectly. MobileNetV2 achieved 97O. Simonoski . [6]O. the spread of COVID-19. The model is cost- effective, can run
Simonoski proposes a real-time and highly accurate approach on low-end computers like Attendance marking, and
for detecting non-detection wearers using a hybrid model of encourages people to wear detections continuously. R.
deep and regular machine learning. The approach uses Sivapriyan
OpenCV to recognize faces in real-time from a live feed . [12]R. Sivapriyan proposed that in today’s fast-paced digital
and sends the data to the cloud for further analysis. An world, a reliable surveillance system is crucial, especially
embedded system utilizing Python, OpenCV, and Tensor with the increased use of video-based surveillance for security
Flow is used to make the process cost-efficient and scalable purposes. This paper proposes an AI-based surveillance
on a Attendance marking. The solution aims to contribute to system that can detect whether a person is wearing a
community health and prevent local transmission of COVID- detection and has a temperature below a specified threshold.
19. The systemuses a cascade classifier generated by Machine
A. Jamini. [7]The COVID-19 pandemic has caused learning to accurately identify detections, despite external
widespread anxiety and disrupted daily life. Wearing factors that can impact traditional surveillance systems.
detections, using hand sanitizer, and monitoring body Attendance marking serves as the central hub for processing
temperature can help prevent transmission of the virus. and analyzing the video stream from the camera.
However, it’s not feasible to manually check everyone’s
compliance. A solution is proposed using a deep learning A. Methodology
system and Attendance marking board to detect individuals Detecting face recognitions using Attendance marking can
not wearing detections or with high body temperature. be accom- plished using a few different methods. Our
Precautions such as a sanitizer dispenser and mobile app alerts approach is to use a camera module connected to the
can then be implemented in crowded areas like schools and Attendance marking to capture images or video of individuals
workplaces. and then use machine learning algorithms to detect whether or
L. Nguyen. [8]This paper proposes an embedded machine not they are wearing a detection. Here are the steps we
learning system for real-time detection of face recognitions followed for approching result in our project
and mea- surement of human temperature using a Raspberry-Pi Setting up Attendance marking with a camera module and
camera and temperature sensor. The system employs a make sure connections are working properly. Install Python
MobileNetV2 model with pre-trained weights to classify and the necessary libraries for machine learning, such as
correctly wearing, incorrectly wearing, and not wearing a OpenCV and TensorFlow. Collection of dataset of images of
detection. The results show that the system achieves high people with and without detectionsis done further Here a pre-
accuracy rates of 98.61 existing
dataset or creating your own by taking pictures of individuals
wearing or not wearing detections can be used . Train a
machine learning model on the dataset. One popular algorithm
for image classification is convolutional neural networks
(CNNs) although a pre-trained CNN like VGG16 or Inception
V3 and fine-tune it for detection detection, model from
scratch. Python program is coded to capture images from the
camera module and use the trained model to detect
detections on individuals in the images. Run the program on
your Attendance marking and tested it out by pointing the
camera at individuals with and without detections. Here is a
flow chart or the work flow for the same: 1 Initialize the
Attendance marking 4 and its camera module. 2 Capture a
video stream from the camera module. 3 Extract individual
frames from the video stream. 4 Preprocess each frame by
resizing it and converting it to grayscale to reduce
computation time. 5 Use a pre-trained face detection model Fig. 1. Example of a figure caption.
to detect any faces in the frame. 6 For each detected face,
use a pre-trained detection model to determine whether the
face is wearing a detection or not. 7 Draw a bounding box Result and discussion
around the face and label it as ”detection” or ”no
detection” based on the detection model’s output. 8 Display
the processed frames with the bounding boxes and labels. 9
Repeat steps 2-8 for each subsequent frame in the video
stream. Optionally, the system could also include a notification
mechanism to alert individuals who are not wearing
detections or send notifications to relevant authorities for
action.

III. PREPARE YOUR PAPER BEFORE STYLING


Before you begin to. [?] format your paper, first write
and save the content as a separate text file. Complete all
content and organizational editing before formatting. Please
note sections III-A–III-E below for more information on
proofreading, spelling and grammar. ADMIN DESK
Keep your text and graphic files separate until after the text
has been formatted and styled. Do not number text heads—
LATEX will do that for you.
• .

REGISTER YOUR FACE


https://ieeexplore.ieee.org/document/9182237#:~:text=In
%20this%20work%2C%20two%2
https://ieeexplore.ieee.org/document/9841195
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4108969
Existing project:
Adrian Rhesa Septian Siswanto, Anto Satriyo Nugroho,
Maulahikmah Gahnium," Implementation of face recognition
algorithm for biometrics based time attendance system", IEEE,
ICT For Smart Society (ICISS), International Conference
January 2015.
Brian C. Becker, Enrique G.Ortiz, "Evaluation of Face
Recognition
Techniques for Application to Facebook" IEEE, 2008.
International Journal of Computer and Communication
Engineering, Vol. 1, No. 2, July 2012 - Study of Implementing
FILL ATTENDANCE
Automated Attendance System Using Face Recognition
Technique by Nirmalya Kar, Mrinal Kanti Debbarma, Ashim
TESTING
Saha, and Dwijen Rudra Pal. Real time face recognition system
using PCA and various distanceclassifiers by Deepesh Raj - IIT
Kanpur

VIEW ATTENDANCE

References
IEEE LINKS

You might also like