Page 473 - Emerging Trends and Innovations in Web-Based Applications and Technologies
P. 473
International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
efficiency but also for ensuring accuracy and reducing Attendance Logging: When a student enters the classroom,
administrative burdens. the camera captures their image, and the system compares it
with stored records in real-time. If a match is found, the
1. Hardware Components
system logs the attendance automatically.
Background and Motivation
The use of biometric systems, including fingerprint scanning, Admin Interface: The system provides an admin panel where
iris recognition, and voice recognition, has been explored in faculty members can view, manage, and analyze attendance
various domains, including educational settings. Among data. Attendance reports are generated in real-time and can
these, face recognition has gained popularity due to its non- be accessed by the administration.
intrusive nature, ease of use, and advancements in artificial
intelligence (AI). Unlike fingerprint or iris scans, face
recognition allows for contactless and remote identification,
reducing the risk of error or fraud.
Despite these advantages, there are concerns related to
privacy, data security, and the accuracy of recognition,
particularly in dynamic environments like a classroom.
FaceAttend was developed to tackle these issues while
automating attendance management.
The FaceAttend System comprises the following hardware
elements:
High-definition cameras positioned strategically in
classrooms or entry points.
A server for data processing and storage.
Optional biometric scanners for supplementary
verification.
The FaceAttend System leverages:
OpenCV for image processing.
Dlib and TensorFlow for facial recognition
functionalities.
A web-based interface for teachers and administrators
to manage attendance records.
Methodology
1. Data Collection
The system initially requires data collection through:
Capturing facial images of students during registration.
Creating a database where these images are stored and
associated with student IDs.
2. Software Components
2. Training the Model
System Design and Architecture
To ensure high accuracy in recognition:
The FaceAttend system consists of the following
The system utilizes convolutional neural networks
components:
(CNNs) trained on the collected dataset.
Data Collection: Students' facial images are captured during Implementation of data augmentation techniques
the registration phase. A high-resolution camera is installed enhances model robustness against diverse lighting and
in the classroom to capture images in real-time during angle conditions.
lectures.
3. Real-time Attendance Tracking
Preprocessing: Captured facial images are preprocessed to The attendance process involves:
normalize lighting conditions and facial positions. This helps Real-time monitoring by facial recognition cameras.
ensure that images are of high quality, which is critical for Automatic logging of attendance as students enter the
accurate recognition. classroom.
Face Recognition Algorithm: A Convolutional Neural System Development Process: Describe the steps taken to
Network (CNN) model, a deep learning architecture, is used develop and implement the FaceAttend system. Include
for face detection and recognition. The model is trained with details about the planning, design, and testing phases.
a large dataset of student facial images, enabling it to learn
Data Collection: Explain how data was collected for the
and distinguish individual features.
purpose of training the face recognition model (e.g., student
Database: The system stores student facial data in a secure images, attendance data, environmental conditions).
cloud-based database, ensuring quick access and backup.
Evaluation Criteria: Outline how the system was evaluated,
Each student is associated with their unique ID and image
such as through accuracy rates, time efficiency, and user
features.
IJTSRD | Special Issue on Emerging Trends and Innovations in Web-Based Applications and Technologies Page 463
