MANGALORE INSTITUTE OF TECHNOLOGY &

ENGINEERING
(A unit of Rajalaxmi Education Trust ®, Mangalore)
Autonomous Institute Affiliated to V.T.U., Belagavi , Approved by AICTE, New
Delhi Accredited by NAAC with A+ Grade & ISO 9001:2015 Certified
Institution
Badaga Mijar, Moodabidri-574225, Karnataka

Course: Mini Project

Course code: BAI586
Project Guide: DR AMRITHAVALLI

Software Requirements Specification (SRS)

Project Title:.: Development of a Face Recognition-Based Automated
Attendance System Using Artificial Intelligence (AI)
Team Members:
Name USN

Mr. Abdul Rehaman 4MT22AI001

Mr. Ihsannul Haque 4MT22AI016

Mr. Mohammed Isham 4MT22AI026

Mr. Ikrammul Haque 4MT22AI017

Problem Statement:.The challenge is to develop an AI-based system that

automatically identifies individuals through face recognition and registers their
attendance without manual intervention. The system should capture real-time
video input, detect faces, compare them with pre-stored profiles, and mark
attendance accordingly. The solution should be efficient, scalable, accurate, and
capable of handling challenges such as lighting variations, facial expressions,
occlusions (e.g., glasses or masks), and different angles.
 Table of Contents
1. Introduction
1.1 Purpose
1.2 Scope
1.3 Definitions, Acronyms, and Abbreviations
1.4 References
1.5 Overview

2. Overall Description
2.1 Product Perspective
2.2 Product Features
2.3 User Characteristics
2.4 Constraints
2.5 Assumptions and Dependencies

3. Specific Requirements
3.1 External Interface Requirements
3.2 Functional Requirements
3.3 System Features
3.4 Non-Functional Requirements

4. Other Non-functional Requirements

4.1 Performance Requirements
4.2 Safety Requirements
4.3 Security Requirements
4.4 Software Quality Attributes
4.5 Business Rules

5. Appendices
5.1 Glossary
5.2 Architecture Diagram
5.3 Use Case Diagram
5.4 Data Flow Diagram (DFD)
 1. Introduction
Purpose
1.1 The purpose of this document is to define the requirements and outline the
design and functionality of a Face Recognition-Based Attendance System
using Artificial Intelligence (AI). This system aims to automate and
streamline attendance tracking in educational institutions, corporate
environments, or event-based scenarios by utilizing face recognition
technology. The system will allow administrators to efficiently track
attendance without manual intervention, reduce errors, and improve security
by using biometric identification methods. The goal is to develop a solution
that is fast, accurate, scalable, and user-friendly.Scope
1.2 Scope

The scope of this project includes the development, testing, and deployment of a Face
Recognition Attendance System that operates using AI models for face detection and
recognition. The system will consist of both hardware and software components, which
include:

1. Face Recognition Algorithms: Algorithms to detect, recognize, and match faces

from a live video feed.
2. Attendance Database: A backend system for storing user profiles, attendance logs,
and time-stamped records.
3. User Registration and Profile Management: A method for enrolling users into the
system, capturing facial images, and associating them with unique identifiers (e.g.,
student ID, employee number).
4. Real-time Processing: A front-end interface where attendance is recorded
automatically as faces are detected and matched to registered profiles.
5. Administrative Interface: A web or desktop application to monitor, modify, and
generate reports on attendance, user profiles, and logs.
6. Data Privacy & Security: The system will adhere to applicable data privacy
regulations (e.g., GDPR) to ensure the secure handling of sensitive biometric data.

The system will be designed to be scalable, capable of handling a small number of users
as well as large organizations with hundreds or thousands of registered faces. The scope
also includes creating an intuitive, easy-to-use interface for administrators and ensuring
the system performs well under various real-world conditions (e.g., lighting, camera
angles)

●
 1.3 References
[1] Schroff, F., Kalenichenko, D., & Philbin, J. (2015). FaceNet: A Unified Embedding
for Face Recognition and Clustering. Conference on Computer Vision and Pattern
Recognition (CVPR).
• Link to Paper
[2] OpenCV Documentation - Open Source Computer Vision Library for real-time
computer vision applications.
• OpenCV Docs
[3] dlib - Dlib is a toolkit containing machine learning algorithms and tools for creating
complex software in C++ to solve real-world problems.
• dlib Library
[4] GDPR Compliance Guidelines - Official European Union website for GDPR
regulations regarding data privacy.
• GDPR Text
[5] TensorFlow - A popular open-source framework for machine learning and deep
learning applications, useful for implementing AI models for face recognition.
• TensorFlow
[6] DeepFace - A Python library for deep learning-based face recognition.
• DeepFace GitHub Repository
[7] Microsoft Azure Face API - A cloud-based API for face detection and recognition
provided by Microsoft Azure.
• Azure Face API
[8] Facial Recognition Technology: The Need for Regulation – A research paper
discussing privacy and ethical concerns in face recognition technology.
• Link to Research

[9]. Smart Attendance Monitoring System (SAMS): A Face Recognition based

Attendance System for Classroom Environment. 2018 IEEE 18th International
Conference on Advanced Learning Technologies.
[10]. A. Kassem , M. Hamad , Z. Chalhoub, and S. EI Dahdaah, An RFID
Attendance and Monitoring System forUniversity Applications.
[11] Hemantkumar Rathod, Yudhisthir Ware, Snehal Sane, Suresh Raulo, Vishal
Pakhare and Imdad A. Rizvi. Automated Attendance System using Machine
Learning Approach. 2017 International Conference on Nascent Technologies in the
Engineering Field (ICNTE2017)
[12]Poornima S1, Sripriya N2, Vijayalakshmi B3,Vishnupriya P4 Attendance
Monitoring System using Facial Recognition with Audio Output and Gender
Classification. IEEE International Conference on Computer, Communication, and
Signal Processing (ICCCSP-2017)
 [13] Rishi Raj, 2Abhinav Das, 3Subhash Chand Gupta . Proposal of an Efficient
Approach to Attendance Monitoring System using Bluetooth
[14] Rashid. R.A., Mahalin, N.H., Sarijari, M.A., and Abdul Aziz, A.A. “Security
system using biometric technology: Design and implementation of voice recognition
system (VRS)”, International Conference on Computer and Communication
Engineering, 2008, pp.898- 902.
 2. Overall Description
2.1 Product Perspective
The purpose of this document is to define the requirements and outline the
design and functionality of a Face Recognition-Based Attendance System
using Artificial Intelligence (AI). This system aims to automate and streamline
attendance tracking in educational institutions, corporate environments, or
event-based scenarios by utilizing face recognition technology. The system will
allow administrators to efficiently track attendance without manual intervention,
reduce errors, and improve security by using biometric identification methods.
The goal is to develop a solution that is fast, accurate, scalable, and
user-friendly
2.2 Product Features

The system provides the following key functionalities:

1. Real-time Face Detection and Recognition:

○ Captures video feeds from a camera and detects faces in the frames.

○ Matches recognized faces with the pre-registered database and marks attendance.

2. User Registration:
○ Users are enrolled by capturing multiple facial images, which are then processed into embeddings

and stored securely in the database.

3. Attendance Logging:
○ Each time a user is recognized, their attendance is marked with a timestamp.

4. Admin Dashboard:
○ A user-friendly interface to view real-time attendance data, manage user profiles, and generate

reports.

5. Data Security and Privacy:

○ The system ensures secure handling of biometric data and complies with relevant privacy

regulations such as GDPR.

6. Scalability:
○ Designed to work efficiently for a small classroom or large organizational settings.

The system provides several benefits, including:

● Increased Efficiency: Automates the entire attendance process.

 ● Accuracy: Minimizes errors associated with manual entry or card-based systems.
● Security: Prevents attendance fraud through biometric verification.
● User Experience: Provides an intuitive and seamless experience for users and administrators.

This solution is ideal for educational institutions, corporate environments, and events where efficient and accurate

attendance management is crucial.

4o mini

You said:

face recognising attendance register using a Product Perspective

ChatGPT said:

ChatGPT

Product Perspective: Face Recognition-Based Attendance Register

Using AI

1. Product Overview

The Face Recognition-Based Attendance System leverages advanced Artificial Intelligence (AI) and

Computer Vision technologies to automate and streamline the process of attendance tracking. The system uses
face detection and recognition algorithms to automatically identify individuals in real-time and mark their

attendance. It operates by processing video input from a camera, comparing the detected face to a pre-registered

database of users, and storing attendance data with an accurate timestamp.

This product aims to replace traditional attendance management methods—such as paper roll calls, manual

sign-ins, and card-based systems—with a faster, more secure, and more efficient biometric solution. It is designed

for educational institutions, corporate offices, events, and other scenarios where tracking attendance is necessary.

2. System Context
The Face Recognition Attendance System will typically interact with the following external entities:

● User Profiles Database: The system will require a database to store the facial data (face embeddings),
along with basic user information such as name, ID, and other relevant details. This database will be used

to match faces detected during attendance with registered profiles.

● Camera/Video Feed: A camera (webcam, CCTV, or mobile device camera) will capture the live video feed
to be processed by the system in real-time. The camera will be responsible for face detection and

recognition.

● Admin Dashboard: The administrator interface will allow administrators to view attendance logs, generate
reports, manage user profiles, and configure system settings.

● User Enrollment: Users (e.g., students, employees) will be required to register in the system by capturing
their face images. These images will be processed into facial embeddings and stored in the database for

future recognition.

● Reporting System: After attendance is captured, the system will allow for the generation of detailed
reports, including attendance status (present/absent), timestamps, and summaries for analysis.

3. Product Features

The product will include the following key features:

1. Face Detection and Recognition:

○ Real-time Detection: The system will continuously process the live video feed to detect faces.
○ Face Matching: Each detected face will be compared to the pre-registered profiles to identify the
individual.

○ Automatic Attendance Marking: Once a match is found, the system will automatically mark the
individual's attendance with the time of detection.

2. User Registration:
○ Facial Data Collection: Users will register by providing multiple images of their faces. These
images will be converted into embeddings (unique face features) and stored in the database.

○ Profile Management: Users' profiles, including face embeddings and personal information, will be
stored securely for future recognition.

3. Attendance Logging:
 ○ Real-Time Attendance Tracking: Attendance will be logged instantly when a user’s face is
recognized.

○ Timestamping: The system will log the date and time of the attendance for each user.
○ Attendance Summary: The system will maintain a record of attendance status (e.g., present,
absent, late) for each individual over time.

4. Admin Dashboard:
○ View Attendance Logs: Admins will be able to view real-time attendance logs, including the list of
recognized individuals and timestamps.

○ Profile Management: Admins can add, update, or remove user profiles from the system.
○ Report Generation: The system will allow admins to generate attendance reports based on time
periods, individual attendance, or group summaries.

○ Alerts: Notifications for missing users, unrecognized faces, or technical issues (e.g., camera
malfunction) will be displayed to the administrator.

5. Data Privacy and Security:

○ Encrypted Storage: Facial data and personal information will be stored securely with encryption to
protect user privacy.

○ Compliance with Data Regulations: The system will comply with relevant data protection
regulations (e.g., GDPR, HIPAA) regarding biometric data usage and storage.

○ Access Control: Access to attendance records and user data will be restricted to authorized
personnel only

●
2.3 User Characteristics
The primary users of the system will include:
● system Configuration and Management: Admins will configure the
system settings, including camera setup, database management, user
profiles, and facial recognition thresholds.
● User Profile Management: Admins will add, modify, or remove user
profiles (e.g., students, employees) in the system. They will ensure that
the data is accurate and up to date.
● Attendance Monitoring: Admins will monitor attendance in real-time,
generating and reviewing reports, and addressing any discrepancies or
issues in attendance data.
 ● Reporting and Data Analytics: Admins will use the system to generate
attendance reports and analyze trends over time (e.g., absenteeism
patterns, attendance performance by department or class).
● Data Privacy Compliance: Admins are responsible for ensuring that the
system is compliant with privacy laws (e.g., GDPR, HIPAA) and that
biometric data is securely stored and processed.
●
2.4 Constraints
 ● Environmental Factors: Face recognition systems rely heavily on good
lighting. Poor lighting, strong backlighting, or shadows can affect the
system’s ability to detect and recognize faces accurately.
● Real-Time Processing: Enrolling users into the system requires capturing
their facial data, which can be time-consuming, especially in large
organizations. Users may also need to re-register if the system fails to
recognize them after a long period.
● Gesture Complexity: Biometric data, including facial recognition
information, is highly sensitive and subject to data protection laws. The
system must ensure that facial data is stored and processed securely.
● Social constraints: AI and facial recognition technologies have been
criticized for biases in recognition accuracy across different demographic
groups, particularly for people of color, women, or individuals with
disabilities. This could lead to unfair treatment or errors in attendance
marking..
2.5 Assumptions and Dependencies
High-Quality Face Data for Registration: It is assumed that users will
provide high-quality face images during the enrollment process, which
will help create accurate face embeddings for later recognition. The face
recognition system assumes that the facial features used for comparison
are clear, unobstructed, and taken in good lighting conditions.

● Implication: If the data quality is poor (e.g., blurry images, poor lighting), the
system’s performance may suffer.

Database Systems: A centralized or cloud-based database is essential to store the facial

embeddings (the mathematical representations of faces) and related user information (e.g.,
names, IDs, attendance logs). This database should be scalable, secure, and capable of
handling large amounts of data.

● Dependency: The availability and performance of a database system (SQL, NoSQL,

cloud-based solutions) for storing and retrieving facial data and attendance records.

3. Specific Requirements
3.1 External Interface Requirements
● Hardware Interface: Cameras capture the face of individuals for
recognition during attendance marking. This can include fixed cameras
 (e.g., mounted on walls) or mobile devices (e.g., smartphones or tablets).

● Software Interface: the system’s face recognition model (which could be

based on deep learning algorithms like FaceNet, DeepFace, or other
proprietary models) requires integration with camera software and
backend systems.
 3.2 Functional Requirements

User Enrollment

● Description: The system must allow users (students, employees) to register their facial data (images) into
the system for future recognition.

● Functional Requirements:
○ Users must take clear and distinct photos for registration.

○ The system should allow for multiple images from different angles for improved recognition

accuracy.

Profile Update

● Description: Users should be able to update their profile if their appearance changes
(e.g., hairstyle, glasses).
Real-Time Face Detection

● Description:
The system must detect a face in front of the camera in real time during
attendance marking.
● requirements:
○ The camera must capture video frames and process them in real time.
○ The system must detect and locate faces in the captured frame.
3.3 System Features
● Real-Time Processing: The system must process facial recognition in
real time to mark attendance without delays.
● Scalability: The system should be scalable to handle different numbers of
users (small classes to large organizations)..
● Camera Frame Rate: The system should work with video frames
captured at 30 frames per second (FPS) or higher to ensure smooth face
detection and recognition, especially when dealing with fast-moving
individuals.
3.4Non-Functional Requirements
● Horizontal Scalability: he system architecture should support
horizontal scaling, allowing it to handle increasing numbers of users
or devices by adding more processing nodes (e.g., servers, virtual
machines) to the infrastructure without significant performance
degradation..
● Usability: the system should notify the user when their attendance is
 successfully marked (via a visual confirmation, message, or sound).
● Reliability: The system must compare detected faces with stored profiles
to identify and verify the user.
 4. Other Non-functional Requirements
4.1 Safety Requirements
Ensure the system complies with data protection regulations like GDPR
(General Data Protection Regulation) or CCPA (California Consumer Privacy
Act). Face images and related biometric data must be handled with care.
4.2 Software Quality Attributes
Accuracy is crucial in a face recognition attendance system to ensure that the
system reliably captures and records attendance without errors. High accuracy
minimizes the occurrence of False Positives (incorrectly identifying someone)
and False Negatives (failing to recognize a valid user).
4.3 Security Requirements
All biometric data (e.g., facial images, facial feature templates) must be
encrypted both in transit (during data transmission) and at rest (while stored in
databases).
5. Appendices
5.1 Glossary

Accuracy

● Definition: The ability of the system to correctly identify or verify individuals. High accuracy minimizes
errors, such as false positives (incorrect identification) and false negatives (failure to identify a legitimate

person).

Algorithm

● Definition: A set of rules or procedures used by the face recognition system to process, analyze, and
match facial data. Common algorithms include deep learning-based models (e.g., Convolutional Neural

Networks - CNNs).

Biometric Template

● Definition: A mathematical representation of a person’s face derived from their facial features. The
template is unique to each individual and is used for comparison during the recognition process instead of

the raw facial image.

Camera Calibration

● Definition: The process of configuring and adjusting the camera to ensure accurate capture of facial
 features. This includes setting optimal angles, focus, and lighting conditions for recognition.

Facial Feature Extraction

● Definition: The process of identifying key facial features (e.g., eyes, nose, mouth, and the distances
between them) from a face image. This data is then used to create a biometric template.

Face Enrollment

● Definition: The process of registering an individual's face into the system. During enrollment, the system
captures a high-quality image of the user's face and converts it into a biometric template for future

recognition.

Face Recognition

● Definition: The process by which the system identifies or verifies a person by comparing their current facial
features (captured in real-time) against stored templates in the system’s database.

Face Matching

● Definition: The comparison between a newly captured facial image or template and stored templates in the
system to find a match, enabling the system to identify or verify the individual.

False Acceptance Rate (FAR)

● Definition: The rate at which the system incorrectly accepts an unauthorized person as a legitimate user.
It’s a measure of the system's security and reliability.

False Rejection Rate (FRR)

● Definition: The rate at which the system fails to recognize a legitimate user (false negative). A high FRR
means that authorized users are frequently denied access.

Liveness Detection

● Definition: A security feature that prevents spoofing attacks by ensuring that the face being scanned is
from a live person and not a photo or video. Techniques might include asking the user to blink, smile, or turn

their head.
 5.2 Architecture diagram: The flowchart outlines a system for recognizing
hand gestures via webcam, converting them to text using CNN, and generating
speech output using pyttsx3 text-to-speech.
User Devices

● Description:
These are the physical devices used by individuals for enrollment and
attendance marking (e.g., smartphones, kiosks, or camera-enabled workstations).
The devices capture facial images or videos for recognition.
● Function: The user's face is scanned and sent to the face recognition module for
identification.
2. Web / Mobile App (User Interface)

● Description:
A user-facing interface that could be a web application or mobile app,
allowing the user to check in (mark attendance) or access their attendance records.
● Function: Provides users with an interface for enrollment, attendance tracking, and
viewing status. It also allows administrators to manage the system, view logs, and
monitor activity.
3. Face Recognition Algorithm Module

● Description:
The heart of the system, where facial images are analyzed and compared
against stored templates using computer vision and machine learning algorithms.
● Function:
○ Face Detection: Identifies a face from the captured image or video stream.
○ Feature Extraction: Extracts unique features of the face (e.g., eye spacing,
nose shape) and converts them into a biometric template.
○ Face Matching: Compares the extracted template with stored templates in
the database to verify identity.
○ Authentication: If a match is found, the system marks attendance; otherwise,
it may trigger an alert or request re-enrollment.
4. User Authentication & Attendance Management

● Description:
This module handles user authentication (via face recognition) and tracks
the attendance data.
● Function:
○ Enrollment: New users are enrolled in the system by capturing their facial
data and storing it as templates.
○ Attendance Logging: Whenever a user’s face is successfully matched, their
attendance is recorded in the system (timestamp, user ID, etc.).
○ Session Management: Tracks active sessions and ensures that each user’s
attendance is recorded only once per session.
5. Secure Database Server

● Description:
A central server that stores biometric templates, attendance records, and
user data securely.
● Function:
 ○ Biometric Template Storage: Stores the facial templates that represent users’
facial features.
○ Attendance Records: Stores logs of attendance events, including the time
and identity of the person marked.
○ Security Features: Data is encrypted at rest and in transit to protect user
privacy and prevent unauthorized access.
6. Database (Attendance & Facial Templates)

● Description:
The backend database that holds all stored data, including:
○ User Profiles: Information such as name, ID, and face template.
○ Attendance Logs: Records of attendance for each user (timestamped).
○ Metadata: Logs for auditing and compliance purposes.
● Function: The database is queried to find matching facial templates, store
attendance logs, and support retrieval for reporting.
7. Notification System

● Description: A system that provides real-time alerts, notifications, and logs.

● Function:
○ Real-time Alerts: Sends alerts to administrators in case of mismatches, failed
attempts, or suspicious activity.
○ Logs & Reports: Generates logs of attendance data, system activity, and any
security events for auditing purposes.
8. Security & Compliance Layer

● Description:
A layer that ensures the system adheres to security policies, encryption
standards, and compliance with privacy laws (e.g., GDPR, CCPA).
● Function:
○ Encryption: Ensures biometric data is encrypted both at rest and in transit.
○ Access Control: Manages access based on roles (e.g., admin, user).
○ Audit Logs: Keeps detailed logs of system access, changes, and events for
compliance and forensics.
○ Compliance Management: Ensures the system meets regulatory
requirements regarding data storage, processing, and consent.
Fig 1. Architecture diagram
5.3 Use Case Diagram: The diagram illustrates a use case where a deaf user
captures ASL gestures, which are processed by the system to detect, convert
gestures to text and speech, and provide both audio and text outputs for deaf and
non-deaf users.

o Input Image: The student face image is acquired using the camera
and the image is acquired from a certain uniform distance the face
of the student should clearly be visible on the screen having a good
resolution of the image with sufficient lighting for learning and
classification.
o Image processing: The data needs to be pre-processed into an
acceptable format for the model. A model does not understand raw
text, images, or audio. The rejection of an acceptable number of
non-face windows is a preprocessing step. Priority is given to
Linear Image Transform (LIT) which ignores scanning several
non-face windows.
o Face Detection: It is closely related to the recognition of human
faces by detecting and recognizes faces in dataset images with
AI-enabled technology. It provides the location of the detected
faces and can perform facial matches to find target subjects.
o Clock In: It makes sures that student logs in to the classroom with
appropriate details like the name of the student, status of the
presence of the student in the classroom and the total number of
hours the student present in the system.
o Clock Out: This is a button where the student can make use of the
button to exit from the classroom and as student logs out as he/she
goes out of the classroom with appropriate details for
documentation.
 Fig 2. Use case diagram

5.4Data Flow Diagram (DFD): This DFD provides a clear understanding of

how data moves through a Face Recognition-Based Attendance System
and how various processes and data stores interact to accomplish the goals of
user authentication and attendance tracking.

1. Face Detection and Feature Extraction:

○ The user’s face is captured by the User Device and processed through Face Detection and
Feature Extraction to create a biometric template.
2. Face Matching and Attendance Logging:
○ The biometric template is then compared against stored templates in the Database for
matching. If a match is found, the Attendance Logging process is triggered to record the
user’s attendance.
3. User Interface:
○ The system’s User Interface provides users or administrators with access to attendance
logs, user data, and system status.
4. Notifications and Alerts:
○ If any issues occur (such as a failed recognition or security anomaly), the Notification &
Alerts process sends real-time notifications to administrators.