Name : Rakesh Basumatary
GU Roll No. : UT-231-014-0032
Department : BCA 6th Semester
PROJECT SYNOPSIS
Project-title : Smart-Attend: Automated Attendance System Using Face Recognition
1. Introduction
In the digital age, the efficiency of administrative processes is paramount for educational
institutions and corporate organizations. Traditional attendance tracking—relying on physical
registers and manual roll calls—is increasingly viewed as obsolete due to its time-consuming
nature and susceptibility to human error. Furthermore, manual systems are highly vulnerable
to "proxy" attendance, where individuals falsely mark presence for absentees.
This project introduces an AI-Based Smart Attendance System designed to modernize this
process. By leveraging Computer Vision and Artificial Intelligence, specifically facial
recognition technology, the system automates attendance tracking. It captures live video,
verifies identity against a pre-existing dataset of facial features, and logs attendance into a
secure database without the need for manual intervention.
2. Problem Statement
The reliance on manual attendance systems creates several operational bottlenecks:
Time Consumption: The roll-call process consumes valuable lecture or working hours.
Data Integrity: Manual records are prone to manipulation and "proxy" attendance.
Human Error: Physical recording and data entry often lead to mistakes.
Inefficient Analysis: Storing and auditing physical records makes generating
attendance reports and analytics difficult.
This project aims to resolve these issues by deploying a contactless, automated solution that
ensures accuracy, speed, and accountability.
3. Objectives
The primary objectives of the proposed system are:
To automate the attendance marking process using Artificial Intelligence.
To eliminate manual recording errors and streamline data management.
To strictly prevent proxy attendance through biometric verification.
To ensure secure, digital storage of attendance history.
To provide an interface for generating comprehensive analytics and reports.
4. Proposed System Architecture
The proposed solution is a web-based application integrated with a robust face recognition
module.
Operational Flow:
1. Capture: The system utilizes a webcam to capture live video frames of the classroom or
entry point.
2. Process: AI algorithms process these frames to detect faces and extract facial
embeddings.
3. Verify: The extracted data is compared against a stored database of student facial
encodings.
4. Log: Upon a successful match, the student is marked as "Present" with a precise
timestamp.
5. System Modules
5.1 Admin Module
The central control panel for the system:
Manage student profiles (Add/Update/Delete).
Capture and encode student facial data for the dataset.
View and audit attendance logs.
Generate daily and monthly attendance reports.
5.2 Face Recognition Module
The core processing engine:
Initiates live video capture via webcam.
Utilizes OpenCV for real-time face detection.
Generates unique facial encodings for identification.
Matches detected faces against the registered dataset.
5.3 Attendance Management Module
Handles data logic and storage:
Records successful matches with Date and Time stamps.
Implements logic to prevent duplicate entries for the same student on the same day.
Facilitates data export to standard formats like Excel or PDF.
5.4 Student Module (Optional)
A user-facing portal allowing students to:
Securely login to verify their status.
View their personal attendance history and aggregate percentage.
6. Technologies Used
Component Technology
Frontend HTML, CSS, JavaScript
Backend Python (Flask Framework)
AI / Computer Vision OpenCV, face_recognition Library
Database MySQL or SQLite
7. Database Design
The system utilizes a relational database with two primary tables: one for storing student
credentials and biometrics, and another for logging attendance events.
Table 1: Student_Master
Stores personal details and facial encoding data.
Field Name Data Type Description
student_id INT (Primary Key) Unique identifier for the student.
name VARCHAR Full name of the student.
roll_no VARCHAR University or Class Roll Number.
department VARCHAR Department or Stream name.
face_encoding TEXT Serialized facial embedding data for matching.
Table 2: Attendance_Log
Stores the daily records of recognized students.
Field Name Data Type Description
attendance_id INT (Primary Key) Unique identifier for the attendance record.
student_id INT (Foreign Key) Links to the Student_Master table.
date DATE Date of the attendance entry.
time TIME Exact time of entry.
status VARCHAR Status (e.g., "Present").
8. System Working (Methodology)
1. Registration: The admin registers a student and captures multiple reference images.
2. Encoding: These images are processed to create a unique "face encoding," which is
stored in the database.
3. Real-time Detection: During a session, the camera activates and detects faces in the
video stream.
4. Matching: The system compares the live face encoding with the stored database.
5. Marking: If the similarity threshold is met, attendance is automatically logged.
9. Advantages & Future Scope
Key Advantages:
Efficiency: Drastically reduces the time spent on roll calls.
Accuracy: Eliminates human errors in data entry.
Security: Prevents proxy attendance via biometric verification.
Analytics: Simplifies the generation of attendance reports.
Future Enhancements:
Cloud Integration: To support multiple classrooms and centralized data.
Mobile App: A dedicated app for parents and students to track status.
Notifications: Real-time SMS/Email alerts to parents regarding absenteeism.
Mask Detection: Integrating safety protocols (e.g., mask compliance) into the
recognition logic.
10. Conclusion
The AI-Based Smart Attendance System represents a significant step forward in
educational administration. By automating a traditionally manual process, it ensures high
reliability, security, and efficiency. This solution not only saves time but also provides a
scalable foundation for modernizing institutional record-keeping.