BUDGE BUDGE INSTITUTE OF TECHNOLOGY

Approved by AICTE, Affiliated to MAKAUT, NAAC Accredited

Nischintapur, Budge Budge, Kolkata -700137, Phone: 033-2482 0676/70 Telefax: 033- 2482 0641
Email: [email protected] , Website: www.bbit.edu.in
NBA Accredited, Valid from AY 2021-22 to 2023-24

Face Recognition-Based Smart Attendance System Using Machine Learning

Submitted in partial fulfilment of the requirement for the award of the degree of
Bachelor of Technology in Information Technology
by

Name University Roll Number

Biplab Mondal 27600222024

Rathin Chandra 27600222022

Tirtho Paul 27600222011

Saheb Mandal 27600222015

Basudev Gorai 27600222014

Under Supervision of Prof. Nivedita Das

TITLE: Face Recognition-Based
Smart Attendance System
Using Machine Learning
SUB TITLE: Revolutionizing Attendance Management
TEAM
Biplab Mondal 1
Rathin Chandra 2
Tirtho Paul 3
Saheb Mandal 4

Basudev Gorai 5
 1 Introduction
Attendance tracking is an essential aspect of maintaining discipline and ensuring
efficiency in both educational institutions and workplaces. It plays a significant role
in monitoring participation, fostering accountability, and enhancing overall
productivity. However, traditional methods of attendance tracking, such as manual
registers and punch cards, are often plagued by limitations including time
inefficiency and human error, which can lead to inaccurate records and increased
administrative workload. To address these challenges, we propose a "Face
Recognition-Based Smart Attendance System Using Machine Learning," developed
using Python, MySQL, and integrated with advanced libraries such as OpenCV and
Dlib. This system leverages machine learning to automate attendance tracking,
offering a more efficient, reliable, and error-free solution.
 2 What is Face Recognition?
A technology that identifies or verifies a person based on their facial features by
analyzing specific characteristics unique to each face.

C
Image Acquisition Capturing the image of a person through cameras or
other devices.
O
M
Locating the face within the captured image and
P Face Detection isolating it for analysis.
O
N Identifying unique facial characteristics such
Feature Extraction as distance between eyes, nose shape, etc., to
E
create a facial signature.
N
Matching the extracted features with stored data
T Recognition and to confirm the person's identity or authenticate
S Verification their presence.
3 Machine Learning in Face Recognition

Role of Machine Learning Techniques Used

• Machine Learning plays a • Neural Networks:

pivotal role in enhancing • A class of algorithms
the accuracy and reliability modeled after the human
of face recognition brain, used to recognize
systems. By using large patterns in facial data.
datasets, these algorithms
learn to identify and • Convolutional Neural
distinguish between Networks (CNNs)
different facial features,
• A specialized type of
continuously improving
neural network that is
their performance over
particularly effective for
time.
analyzing visual data like
images, by capturing
spatial hierarchies in
facial features.
4 What We Need ……….?
Components Hardware
• Camera Module
• Cameras
• Captures images in real-
time, providing input for • Server or Cloud
face recognition. Services
• Processing Unit Software
System Architecture

Technology Stack
• Runs the machine
learning models (e.g., • Programming
using CNNs) to detect and Languages
recognize faces from the
captured images.
• Python
• Database (MySQL) • Java
• Stores user profiles, facial Libraries
data, and attendance
records securely. • OpenCV
• User Interface • TensorFlow, Keras
• An admin dashboard
where attendance records Database
can be viewed, managed, Management
and exported, providing a
seamless interaction for • MySQL
administrators. • MongoDB
 5 Working Mechanism…….!

Image Capture Face Detection Feature Face Attendance

•Cameras capture •Algorithms like Extraction Recognition Marking
images of Haar Cascades or •Convolutional •The system •Once recognized,
individuals during Deep Neural Neural Networks compares the the system
class or work hours Networks (DNN) are (CNNs) extract detected faces with automatically
in real-time. used to detect and unique facial stored profiles to updates attendance
locate faces within features using facial identify or verify the records in the
the captured embeddings to individuals. database without
images. represent the any manual
distinct intervention.
characteristics of
each face.
 6 Training the Model……..!
Train the machine
Gather diverse learning model
images of using labeled data,
individuals, where each image
Augment data
capturing various is paired with the
(e.g., rotate, flip)
angles, lighting Resize images to corresponding
Normalize images to increase
conditions, and fit model identity. This
for consistency. dataset variety
expressions to requirements. allows the model
and improve model
create a to learn and
robustness.
comprehensive accurately
dataset for distinguish
training. between different
individuals.

Data Model
Preprocessing
Collection Training
 7 Challenges and Solutions……..!
Environmental Privacy
Variability:
Factors: Concerns:

Challenge Challenge Challenge

•Variations in lighting and •Collecting and storing •Differences in facial
camera quality can affect facial data raises privacy expressions, angles, and
recognition accuracy. issues. appearances can reduce
model performance.

Solution Solution Solution

•Use high-quality cameras •Implement strong data •Apply data augmentation
and optimize lighting encryption and ensure techniques to diversify
conditions to improve compliance with privacy the dataset, making the
detection and regulations (e.g., GDPR) model more robust to
recognition. to safeguard user data. variability.
 8 Advantages……….!

Accuracy Efficiency Scalability Data Insights

• Machine learning • Automated • The system can • Provides valuable
algorithms attendance easily be scaled analytics and
ensure high marking is fast, to accommodate insights into
precision in face reducing time larger groups, attendance
recognition, spent on manual whether in patterns, helping
minimizing record-keeping. classrooms or organizations
errors. workplaces. make informed
decisions.
9 Future Enhancements……..!

Integration
•Link the system with existing
school management systems
or HR tools for seamless data
synchronization and enhanced
functionality.

Continuous
Learning Mobile Access
•Implement feedback •Develop a mobile app to
loops to continuously allow administrators,
retrain and improve the teachers, or HR
model, ensuring that it personnel to track
adapts to new data and attendance and access
evolving facial reports remotely.
characteristics.
 10 Conclusion…………!
• The Face Recognition-Based • This innovative system is a step
Smart Attendance System forward in improving attendance
provides an accurate, efficient, management. Stakeholders are
and scalable solution for encouraged to consider
automating attendance implementing this solution to
management. By leveraging enhance efficiency, streamline
machine learning algorithms, it processes, and stay ahead with
minimizes human error and saves modern technology.
time, while offering valuable data
insights and analytics.

Call to
Summary
Action
 11 Sample Of Our Project
For For New
already Registration
Registered
Employee

Registered
Employee
make New Employee 1st
attendance by Enter their ID &
pressing “Take Password then
Attendance” press “Take
Button Images” button
12 Sample Of Our Project

After click on take

images button, it
will take 100 of
images of the
human face only for
better processing.
 13 Sample Of Our Project
After take images
click on “Save
Profile” button

To save the
Profile it’s
need to enter
“password”
which is set by
the “Admin”
 14 Sample Of Our Project

After that it will open the camera and

Already registered Employee just click
scan the human face if the face already
on “take attendance” button
registered then showing the name of the
employee and make the attendance
15 Sample Of Our Project
This Attendance
will update day by
day and showing the
information here.
 16 Sample Of Our Project

Attendance list also available in excel

This is the list of registered employee.
format, and it will also update day by day
Thank
you